A COMPARISON OF THE EFFECTS OF TWO APPROACHES
TOWARDS PRONUNCIATION INSTRUCTION INVOLVING
TWO GROUPS OF BEGINNING LEARNERS OF ENGLISH AS A
FOREIGN LANGUAGE
by
NILTON VARELA HITOTUZI
A dissertation submitted to the
School of Humanities
of the University of Birmingham
in part fulfilment of the requirements
for the degree of
Master of Arts
in
Teaching English as a Foreign or Second Language (TEFL/TESL)
This dissertation consists of approximately 12,116 words
Supervisor: Elizabeth Harrison
Centre for English Language Studies
School of English
University of Birmingham
Edgbaston
Birmingham B15 2TT
England
March 2004
ABSTRACT
The study reported in this paper aimed to experiment with two different approaches
towards pronunciation teaching: one based solely on mimicry and implicit learning
through exposure to the speech of the native-speaker parameter on an audiotape, and the
other capitalising on the awareness of phonemic symbols and instances of the discourse
intonation approach. Objectively the investigation sought to discover which of two
groups of beginning learners of English as a foreign language (EFL), having received as
little target-language input as possible, and with little or no familiarity whatsoever with
the data involved in the experiment, would outperform the other in terms of accuracy
and intelligibility after being given different treatments on pronunciation. At first it was
hypothesised that, even having little exposure to a native-speaker model, the group
receiving a treatment based on segmental and suprasegmental phonology would
outperform the group undergoing a treatment basically involving exposure to a nativespeaker model and repetition drills. The results indicate that the use of either approach
can render similar levels of accuracy and intelligibility of EFL beginning-learner
pronunciation production. Nevertheless, this conclusion is compromised by the
conditions under which the experiment was conducted and the inconsistency of the
raters assessment of the pre- and post-tests administered to the groups involved.
Finally, it is argued in this paper that, possibly, the reproduction of this experiment in a
language school, where apparently the clientele is more motivated to learning the
foreign target language than that in a state school (the type of school in which the
experiment was conducted), for example, would yield more conclusive results.
DEDICATION
To God, my creator.
ACKNOWLEDGEMENTS
I would like to thank Professor Elizabeth Harrison, my supervisor, for her thoughtful
comments on this dissertation. Additionally, particular thanks are due to Mr. Walquimar
Vilaça Batista Borges, the headmaster of the state school where the experiment reported here
was conducted, and to the students who volunteered for the study. Likewise, I am indebted to
my wife, who has always been supportive. I would also like to thank Mrs. Olga Bastos, the
coordinator of PECLE (Projeto de Ensino com Capacitação em Línguas Estrangeiras), for
her help and continued support.
CONTENTS
CHAPTER 1
INTRODUCTION
1
1.1
The rationale for the study
1
1.2
The aim, objective and some limitations of the study
4
CHAPTER 2
SOME CONSIDERATIONS ON FORMAL PRONUNCIATION
TEACHING AND SOME OF THE APPROACHES ADOPTED
TO IT
6
2.1
The role of pronunciation in the second-language classroom
6
2.2
On the meaning of prosodic elements
11
2.3
Meaning through intonation
13
2.3.1 Some features of the intonation system
13
2.3.2 The construct of pitch
14
2.3.3 The significance of prominence
15
CHAPTER 3
THE EXPERIMENT
23
3.1
The research question
23
3.2
The research hypotheses
23
3.3
Subjects
23
3.4
Material
24
3.4.1 Worksheets used with the Control Group
25
3.4.2 Worksheets used with the Experimental Group
25
3.5
26
Procedure
CHAPTER 4
THE ANALYSIS OF THE QUESTIONNAIRE,
THE PRE- AND POST-TESTS
28
4.1 The questionnaire
28
4.2 Independent raters assessment of the pre- and post-tests
31
4.3 The statistical analysis of the tests
33
4.3.1 The pre-test
34
4.3.2 The post-test
36
4.3.3 A within-groups comparison of the sample mean scores
38
4.4 Comparing the NSP systems of intonation with those of CG and
EG members
CHAPTER 5
40
THE RESULTS REVISITED AND SOME
CONCLUDING REMARKS
45
5.1
Factors related to the context of the investigation
45
5.2
Factors related to the assessment of the pre- and post-tests
46
5.3
The NSP DI instances used in the speeches of the CG and EG members
48
APPENDIX I: The questionnaire
50
APPENDIX II: The reading test
51
APPENDIX III: Worksheets A to G
52
APPENDIX IV: The flashcards
59
APPENDIX V: Tables 1 to 5
62
APPENDIX VI: Structure of the use of the intonation systems of prominence,
tone and key by CG and EG members
67
REFERENCES
72
LIST OF TABLES AND FIGURES
Dimensions of spoken communication
System of pitch movement
The key system
Table 1
13
Table 2
18
Table 3
The termination system
Critical Values of t
19
Table 4
21
Table 5
66
Assessment of the pre-test administered to CG and EG members
Assessment of CG and EG post-tests
Table 6
35
Table 7
37
Comparison of within-groups means: pre- and post-tests Table 8
The frequency of DI instances in the speech of 8 imaginary NSPs
39
Table 9
Frequencies of NSP DI instances in the speeches of CG and EG members
CG and EG monthly exposure (in hours) to spoken English
Rate of candidates attitudes towards learning English
Respondents attitudes towards pronunciation accuracy
Figure 1
Figure 2
Figure 3
Estimate of the correlation between inter-raters' scores on the post-test
for the experimental and control groups Figure 4
42
Table 10
44
29
30
31
33
CHAPTER 1
INTRODUCTION
'The value of a clear and intelligible pronunciation for the purposes
of communication should be understood by all learners.'
H. H. Stern
1.2
The rationale for the study
A basic question underlying this study is whether or not pursuing the achievement of
good pronunciation of a foreign language (FL) through formal instruction is worth the
time and effort. One of the arguments in favour of the value of good pronunciation may
lie in the concept of self-representation.
Ivani and Camps (2001) maintain that the concept of self-representation is ever present
in people s behaviour; and that it is filtered through various nonverbal and verbal
semiotic resources. The former can be conveyed by the manner in which one sits, eats,
dresses or walks. If, for instance, one makes noise during mastication in a restaurant in
Brazil, they will certainly be associated with a pig, which is extremely negative in this
country. The accent with which people speak is an example of the latter. The desire of
being identified as a Western African expressed by the Nigerian woman in Ivani and
Camps (2001: 5) seems to be a classic example of how verbal semiotic resources can be
used to cast targeted self-representation.
Apparently, self-representation can be circumstantially voluntary or involuntary, though.
At times one may choose to be viewed as a ruthless or kind person, for instance, for any
particular reasons. At other times, one may be perceived as such, as a result of the
working of other people s psyche (Martins, 1982). It seems that the choice (or
imposition) of self-representation can be determined by the same influences that can
trigger motivation. Shedivy (2004) comprises them into two major categories, namely
(1) integrative and (2) instrumental orientations. These orientations can be viewed as the
desire to conform to sociocultural and socio-political values, and the pursuit of personal
economic interest. Such motivational factors can be influential in the way people s selfimage is cast and the choice of their career. A prime example of this is the experience of
1
Cindy Jackson (2004), who has had a successful career as a cosmetic surgery adviser
after having undergone several cosmetic surgeries in search of stereotyped female
beauty.
Having said that, it is claimed in this paper that pronunciation too is a source of selfrepresentation, and as such, it can have either positive or negative effects on other
people, which, in turn, reflect on the speaker as a sign of success or failure in his or her
professional or social (or both) dealings.
Those who learn the spoken form of a foreign or second language with an instrumental
orientation in mind (Gardner, 1985, cited in Shedivy, 2004), for instance, will most
probably seek to pronounce it as intelligibly and as close to the native parameter adopted
as possible (Koren, 1995: 387). For these learners, good pronunciation is a major
element for success in their careers.
Bad pronunciation, on the other hand, seems to jeopardise people s opportunities both in
the professional and the social sphere. What student whose objective was to become, for
instance, a TV anchor in a Spanish-speaking country would like to have someone who
spoke Portunhol (a mixture of Portuguese and Spanish) as their Spanish teacher? Or
how long would someone bear socialising with people who can hardly understand what
other people say or be understood by them? In this same vein, Poedjosoedarmo (2004: 8)
posits that fluent and well-educated speakers usually associate bad pronunciation with
less prestigious accents . This point is further strengthened by Graham:
if one s written or spoken errors place one s apparent performance in the comic
or villainous zones of perception, the message will be devalued in the minds of
the recipients.
Graham (nd)
Probably, this is one of the reasons why Gardner (1985, cited in Koren, 1995: 388)
suggests that pronunciation is pivotal for those having integrative motivations.
2
The findings of a study conducted by Bresnahan et al. (2002) involving undergraduates
in the United States of America further exemplify the effects of pronunciation on other
people. They indicate that the majority of the subjects exhibited negative attitudinal and
affective response towards foreign teaching assistants who had a strong foreign accent
(low level of intelligibility). Foreign friends and teaching assistants displaying high level
of intelligibility, on the other hand, were perceived as being attractive and competent
(Bresnahan et al., 2002: 182). Similarly, in a study on cultural and situational influences
on FL learners beliefs about language learning,
Horwitz (1999) reports that the
majority of the respondents to the Beliefs about Language Learning Inventory (BALLI),
in seven studies analysed by her, considered important to speak an FL with an excellent
accent (Horwitz, 1999: 566).
Carruthers (1987: 192) claims that not having a good pronunciation can sometimes be a
positive factor. This is most comforting for those who, for various reasons, cannot or
would not touch the threshold of what can be considered good pronunciation. But this
does not change the fact that bad pronunciation can ruin one s chances in various fronts;
ultimately it can even cause communication breakdown. Koren seems to endorse this
stance. She makes the point that [good pronunciation] is part and parcel of successful
communication (Koren, 1995: 388). Moreover, Goh (2000: 62) points out that the
inaccurate pronunciation imprinted on ones
mind is of little, if any use for
understanding accurate pronunciation, which means to say that there is a close link
between production and perception of pronunciation (Kello, 2003: 620). This is likely to
be another reason why L2 learners should focus on accuracy and intelligibility when
learning an L2 for oral communication purposes.
Language teachers concurring on this view will presumably have their pupils work on
pronunciation right from an elementary level. Nevertheless, perhaps many of them,
capitalising on suggestions from coursebooks, or approaches with no empirical
authority, might produce speakers with serious pronunciation problems in the target
language. Thus, concerned with their pupils development as they may be, it is almost
certain that, reflecting on their students pronunciation difficulties, they must wonder
how they should approach formal pronunciation teaching (FPT).
3
Sharing this concern with pronunciation quality, I, too, keep wondering what approach
(or approaches) and techniques I should use in my lessons so that my pupils
pronunciation can be enhanced. In my observations of lessons given by other EFL
teachers, I have noticed a constant pattern on their approach towards pronunciation
teaching: (1) mimicry of a native-speaker parameter on tape, and (2) exposure to the
spoken language through audio/video materials
perhaps as an attempt at helping the
learners acquire pronunciation implicitly (Krashen, 1983 and elsewhere). However,
when comparing this trend with the approach adopted in ELT materials, such as the
English File (Oxenden and Seligson, 1996) and the Hotline (Hutchinson, 1998) series,
and Underhill s (1994) Sound Foundations, it becomes evident that this pattern is by no
means the only current approach towards pronunciation teaching.
I for one, as an EFL learner myself, can only feel confidence in uttering a new word
after reading its phonemic transcription in a dictionary. On might claim that this is a
default way of leaning pronunciation in the absence of, or for not exploring, other
techniques; or that this may be a habit. Whether one thing or the other, the truth is that
segmental phonology (Cauldwell and Allan, 1998) has given me a sense of selfconfidence when using English in oral communication. Thus no further empirical
experiment would be needed to lend credence to the usefulness of phonemic symbols for
enhancing pronunciation quality and self-confidence when verbalising English words
were not for the fact that individuals learning styles must be considered when making
generalisations about the effectiveness of teaching approaches (Gregorc, 1984; Dunn et
al., 1995; Frederico, 1991, all cited in Miller, 2004).
1.2
The aim, objective and some limitations of the study
The awareness that there is more to FPT than imitation activities and exposure to TL
models, combined with the desire to disseminate rudimentary notions of Brazil s (1994
a, b and elsewhere) Discourse Intonation (DI) approach amongst EFL beginning
learners, and my own EFL learning experience were the elements that triggered the
design and implementation of the experiment described in this paper.
4
Basically the aim of this study was to experiment with two different approaches towards
FPT: one based solely on mimicry and implicit learning through exposure to the speech
of the native-speaker parameter (NSP) on an audiotape, and the other capitalising on the
awareness of phonemic symbols and instances of the DI approach. The experimentation
involving this approach was an attempt at introducing basic notions about the division of
speech into tone units, and about the intonation systems of prominence, tone and key to
beginning learners of English. The fact that Brazil s (1994a, b) course on pronunciation
was designed to cater for advanced learners needs and the lack of evidence of similar
attempt in the literature seem to indicate that this investigation had a pioneering element.
The objective of the experiment was to verify which of two groups of EFL beginning
learners, having received as little TL input as possible, and with little or no familiarity
whatsoever with the data involved in the experiment, would outperform the other in
terms of accuracy and intelligibility after being given different treatments on
pronunciation.
The initial hypothesis was that, even having little exposure to an NSP, the group
receiving the treatment based on segmental and suprasegmental phonology (Cauldwell
and Allan, 1998) would outperform the group undergoing a treatment basically
involving exposure to an NSP and repetition drills throughout the sessions.
Finally, it is also imperative to note some important caveats. Firstly, the investigation
was conducted by a non-native-speaker-of-English novice teacher-researcher. Secondly,
the fact that the experiment was conducted in a school where English is studied as a
compulsory subject can be considered a major variable in the study. Thirdly, the two
available groups for the experiment were composed by teenagers whose attendance at
the experiment sessions was rather irregular. Another factor may be that the gaps
between sessions seem to have posed another intervening variable to the experiment.
Finally, and importantly, the unavailability of experts in the field of phonetics and
phonology to assess the tests administered to the participants in the study was also a
major setback. Thus, in the light of these limitations, the results presented here are
tentative at best.
5
CHAPTER 2
SOME CONSIDERATIONS ON FORMAL PRONUNCIATION TEACHING AND
SOME OF THE APPROACHES ADOPTED TO IT
2.1
The role of pronunciation in the second-language classroom
Although pronunciation has been taught formally in the L2 classroom from time
immemorial, it remains a controversial issue for the variety of stances addressing it. A
diachronic analysis, as starting from the 1900s, reveals interesting facts about the way
pronunciation has been approached by language teachers along the years.
FPT seems to group language-oriented writers and teachers into at least three broad
categories: (1) those who disapprove of giving pronunciation special attention in the
classroom, adopting thus an intuitive approach to it; (2) those who, even favouring an
analytic approach towards pronunciation, draw heavily mostly on imitation techniques
involving repetition drills based on native-speaker models; and (3) those who believe
awareness of the sound system at both the segmental and suprasegmental levels with the
help of visual aids
e.g. phonemic charts (Underhill, 1994), and arrows to indicate pitch
movement (Brazil, 1994a, 1994b)
is also an important tool for enhancing accuracy
and intelligibility of FL learners pronunciation production.
Defenders of the Classical Method or the Grammar-Translation Method (GTM), as it
was called later in the early 1900s, allotted little or no role to pronunciation in the L2
classroom. It is a GMT tenet that written language is superior to spoken language, thus
the focus of a GTM lesson is essentially on grammar, vocabulary, meaning and mental
exercise through memorisation aiming at developing learners minds (Larsen-Freeman,
1986: 9-11).
On the rise of the twentieth century, proponents of the Direct Method (DM), in a
reaction against the GMT, argued that learning a language was synonymous with being
able to speak it intelligibly, and that the leaning process of an L2 would be the same as
that of learners L1. The L2 classroom should therefore capitalize initially on the spoken
language. Apparently, in a DM lesson, pronunciation teaching is emphasized through
6
learner exposure to the target language repetition drills, and noticing of individual
sounds (Larsen-Freeman, 1986: 20-21).
In the 1950 s, borrowing the basic tenets of the DM, Fries (1945) and other languageoriented theorists founded the Audiolingual Method (ALM)
first known as the Army
Method (Brown, 1994: 57). One aim of the ALM founders was to produce native-likespeaking learners through the Army Specialised Training Programme (ASTP) to meet
the demand of the US Army, which needed fluent language speakers during the World
War II. Drawing heavily on Bloomfield (1933), Pavlov (1955) and Skinner (1957), they
established that pronunciation should be taught mostly through repetition drills for
conditioning and habit formation (Newton, 1979: 18).
In the last quarter of the twentieth century, after the euphoria about the DM and ALM
had receded, once more L2 writers began playing down the role of formal pronunciation
teaching in the classroom. The new argument now was that pronunciation was learnt
intuitively, thus L2 teachers could dispense with FPT. Amongst the most fervent
advocates of the intuitive approach, in this post-reform era, are Krashen and Terrell
(1983). They make a strong case against early pronunciation practice: In the Natural
Approach we do not recommend any specific activities for pronunciation, especially in
the early stages (1983:90). Krashen and Terrell are sceptical of the belief that early
pronunciation practice has much effect on ultimate pronunciation. Instead they believe
that exposure to the target language alone suffices for the acquisition of intelligible
pronunciation. Additionally, Krashen and Terrell dismiss as unfounded the argument
advanced by some language-oriented professionals that premature fossilisation may be a
corollary of little or no emphasis whatsoever on pronunciation in the L2 classroom.
While polarising with not a few language-oriented writers (Callamand, 1981; Canale
and Swain, 1980; Canale, 1983; Callamand and Pedoya, 1984; Celce-Murcia, 1987;
Naiman, 1992, all cited in Stern, 1992; see also Long, 1983), these authors stance
concerning FPT appears to be endorsed by other expert views. One such is Broughton
et al (1978) who half a decade before them had already hinted their allegiance to the
intuitive approach towards L2 pronunciation while writing on how to approach FPT
(emphasis added):
7
Pronunciation teaching deals with two interrelated skills
recognition or
understanding the flow of speech, and production or fluency in the spoken
language. These skills rely very little on intellectual mastery of any pronunciation
rules. Ultimately it is only practice in listening and speaking which will give the
learner the skills he requires.
(Broughton et al, 1980: 49)
This seems to be in line with Newmark s (1990) views. Besides dismissing the teaching
of pronunciation through phonetics as counterproductive, this author suggests that
through close observation one can learn to use quasi-native pronunciation in speaking
the target language.
Leather (1983) also seems to share Krashen and Terrell s view. This author claims that
excessive emphasis on correct pronunciation may prevent learners from learning
grammar or vocabulary. A possible answer to this concern of Leather s appears to be the
implementation of activities in the L2 classroom aiming to raise learner awareness
(Rutherford, 1987), stressing the importance of intelligible pronunciation vis-à-vis other
aspects of the target language (e.g. grammar and vocabulary), since the interdependence
of such components seems to be pivotal to achieving successful oral communication
(Koren, 1995: 388). Furthermore, Leather posits that attention to pronunciation should
be given based on the specific objectives the L2 programme. Considering the variety of
actual L2 learner needs, this seems to be an obvious suggestion, though, inasmuch as
there are certain L2 programmes which focus solely on reading text comprehension, for
instance. In such programmes, less work on pronunciation might be required, as Stern
points out:
In our view, then, pronunciation is never unimportant, although the level of
accuracy of pronunciation as an objective may well vary considerably for different
types of courses (Stern, 1992: 116).
.
Despite focusing on a limited number of phonological items, Nunan appears to be yet
another supporter of the intuitive approach towards pronunciation. Following Sato s
8
(1985) suggestion that some phonological features might be instruction proof, he
maintains that the production of these features cannot and should not be forced by
instruction (Nunan, 1991: 105; see also Scarcella and Oxford, 1994), implying either
that there are language items (LI) that can only be learnt naturally, or that attempts at
teaching LI which are beyond learners current processing level are doomed to failure,
as claimed by supporters of the learnability and teachability theory (Pienemann, 1984).
Nevertheless, FPT seems to have grown in prominence again, this time in various
guises, yet not without overlapping boundaries. The primary example of this is
Carruthers (1987: 191-199), who, adopting a rather eclectic approach, makes the point
that pronunciation practice should be incorporated into the lesson through a variety of
techniques, including work on minimal pairs, and use of articulatory charts, tongue
twisters, limericks and backwards build-ups. He also favours the use of a phonemic
script as an aid to teaching adults pronunciation.
Some advocates of the Communicative Teaching Approach (CTA), for example, while
admitting the necessity for FPT in the L2 classroom, have tried to establish a connection
between it and meaningful communicative practice (Pica, 1984, 2000; Pennington and
Richards, 1986; Wong, 1987; Pennington, 1996). This seems to be an interesting way of
approaching pronunciation in the L2 classroom, since the dearth of connection with
anything realistic to learners which learning (the written or the spoken form of) a
language in isolation generates (irrespective of learner proficiency level) appears to
make learning a more difficult task. Insofar as pronunciation is concerned, a concurrent
view is expressed by Stern:
At no stage in a pronunciation syllabus should learners lose sight of the overall
picture. They cannot deal successfully with particular sound segments without
being able to relate them regularly to the way these sounds manifest themselves in
the stream of speech. This fact should always be borne in mind in designing a
pronunciation syllabus.
(Stern, 1992: 116)
The importance of context is further echoed in Ohala and Shriberg s (1990, cited in
Francis and Jones, 1996: 388) experiment on sound perception. They found that subjects
had difficulty to identify target vowels which had been severely low-pass filtered to
9
eliminate frequency information above 1000 Hz. when these vowels were heard within
the context of a sentence which had not been filtered in the same range. The vowels
were reliably identified, though, when the context sentence was equally filtered (see also
Firth, 1957; Sadock, 1978, both quoted in Brown and Yule, 1983: 37).
The visual and kinaesthetic approach towards FPT also clearly exemplifies the diversity
in approaches and the growing concern with pronunciation in the L2 classroom (Acton,
1984; Gilbert, 1993; Brazil, 1994a, b; Pennington, 1996). This approach is part of a
broader view of pronunciation teaching which advocates the use of multisensory modes
as major tools to cater for different learning styles. Wrembel (2001: 65) describes four
types of multisensory reinforcements which are being applied successfully by many
pronunciation practitioners to make their lessons more effective : (a) visual (e.g.
phonemic charts, diagrams and flashcards; (b) auditory (e.g. repetition drills and
memory pegs ); (c) tactile (e.g. finger tips to feel vibration of the vocal chords, elastic
bands to illustrate vowel length, and a piece of paper to introduce aspiration); and
kinaesthetic (e.g. tracing intonation contours with arms, modelling the mouth with
hands, counting the number of syllables on fingers, clapping or stamping the rhythm ).
Within this same vein, however, there are some polarising views. While some insist on
sound discrimination (O Connor and Fletcher, 1989), others make a case against it,
claiming that perception may not precede production (Goto, 1971). Yet, others seem to
take a more harmonising stand by favouring concomitant work on listening and
repeating (Leather and James, 1991; Pennington, 1996; Gilbert, 1993, Rogerson and
Gilbert, 1990).
Underhill's (1994) Sound Foundations is exemplary of this line.
Adopting a pragmatic, bottom-up approach, through his three-level discovery and
classroom toolkits, he systematically introduces segmental phonology as an attempt at
making his readers perceive it through their senses.
Resorting to Contrastive Analysis (CA) as an attempt to tackle L2 learner pronunciation
difficulties is not uncommon, mainly in university courses on phonetics and phonology
(Baker, 1977; Kenworthy, 1987; O Connor and Fletcher, 1989; Bowler and
Cunningham,
1991;
Deterding
and
10
Poedjosoedarmo,
1998).
Tarone
(1978),
nevertheless, disapproves of an approach drawing heavily on CA. This author sustains
that interlanguage (IL) phonology is affected by other more important factors, such as
avoidance, overgeneralization and approximation (see also Flege, 1987; Eckman, 1977;
Maken and Ferguson, 1987). Such dismissal of CA as tool for pronunciation teaching
has not discouraged other language-oriented writers and teachers of advocating its
usefulness in the L2 classroom. Gloria Poedjosoedarmo, for instance, in O Ensino da
Pronínucia: por quê, o quê, quando e como (The Teaching of Pronunciation: why, what,
when and how
my translation), which has been recently published, suggests sound
contrast activities for improving pronunciation accuracy (Poedjosoedarmo, 2004).
More recently, there has been a great shift towards suprasegmentals triggered by the
growing sensitivity to the communicability of prosodic features in spoken discourse
(Brazil et al, 1980; Sinclair & Brazil, 1982; Brown and Yule, 1983; Brazil, 1985;
Coulthard, 1985; McCarthy, 1991; Brazil, 1994a, b; Gilbert, 1993). A good example of
this is Discourse Intonation (DI), the approach devised by David Brazil at the University
of Birmingham, which has added a new and attractive chapter to the history of
pronunciation teaching.
Koren (1995:389) maintains that suprasegmentals play a major role in the
developmental process of L2 learner IL phonology. This is an area which deserves
further elaboration in this paper for the fact that intonation was also subject to analysis
in the study reported here. Thus an elaboration on the meaning of prosodic features and
the intonation system of prominence, tone, key and termination as proposed by Brazil
will be conducted in the next two sections.
2.2
On the meaning of prosodic features
Apparently, the interpretation of a given message is dependent on a series of interrelated
conditions. Firstly, a crucial aspect for that appears to be the profiles, or the
interpenetrating biographies (Coulthard, 1985:106) of the participants
writer/speaker
on the one end of the channel, and reader/hearer on the other. Thus, the way the
participants perceive the world and the socio-cultural relationships involved in it can
determine how they interpret a given communication. Schank (1979: 400, cited in
11
Brown & Yule, 1983: 207) strongly concurs with this view: humans understand what is
said to them in terms of their own knowledge and beliefs about the world .
Secondly, as it has been previously mentioned, context appears to be another key
element for understanding both spoken and written texts. Firth (1957: 226; see also
Sadock, 1978: 281, both quoted in Brown and Yule, 1983: 37) endorses this point by
suggesting that situational contexts play an important role in understanding verbal
exchanges. The idea that context variation may be tantamount to meaning variation
seems to be a tenable one (Fillmore, 1977). Consider the meaning of the word storm the
following utterance in two different contexts:
The storm began five minutes after the politician had started his speech.
Context 1: A politician is delivering a public speech in a winter season in a city in the
Amazon rainforest. In this scenario, storm may be synonymous with a violent
disturbance of the atmosphere with strong winds and
with thunder and rain
(The
Concise Oxford Dictionary of Current English, 1990:1202).
Context 2: A presidential candidate in Colombia in the 1990 s outlining his plans for
dealing with the drug cartel in that country in a public appearance. At this juncture it is
likely that the word storm is being employed as an umbrella term to describe the
disturbance resultant of an assassination attempt perpetrated by drug dealers.
Another interesting example of this can be found in Brown and Yule (1983: 36; see also
Coulthard, 1985: 2).
Finally, Brown and Yule (1983:209) make the point that intonation can be yet another
non-formal meaning constrainer. Nevertheless, the predictability of meanings through
this suprasentential substance seems to be highly intractable for the discourse analyst,
especially because the construction of meaning based on it seems to be realised on a
moment-to-moment basis by the speaker in real-time interactions. Perhaps this is why
discourse analysts would rather dissect teacher talk than conversation outside the
12
classroom (Sinclair & Brazil, 1982; McCarthy, 1991; Coulthard, 1985).
The next
section reviews the contribution of intonation towards an understanding of spoken texts.
2.3
Meaning through intonation
2.3.1 Some features of the intonation system: It seems common ground amongst many
linguists that intonation also plays a major role in determining the meaning of linguistic
items. However, before exploring their views, it is deemed necessary to briefly outline
some of the main features of the intonation system.
According to Cauldwell and Allan, as demonstrated in Table 1, a reproduction from
Couper-Kuhlen (1986:7, cited in Cauldwell and Allan, 1998: 4), there are three basic
dimensions involved in the oral/aural communication system.
Table 1
DIMENSIONS OF SPOKEN COMMUNICATION
Articulatory
Acoustic
Auditory
Vibration of vocal
folds
Fundamental frequency
Pitch
Physical effort
Amplitude (intensity)
Loudness
Timing of movements
Time
Duration
Cauldwell and Allan (1998: 4-5) also point out that the perception of pitch, loudness, and
duration, all in the auditory column, is resultant of the interaction of the three dimensions
in question. Moreover, they go on to say that these prosodic features
are
the
terms
through which intonation is commonly studied. Nevertheless, perhaps due to the daunting
intricacy of the matter, many discourse analysts have chosen to describe intonation in
terms of pitch alone (Coulthard, 1985:101).
13
2.3.2 The construct of pitch:
McCarthy (1991: 99) maintains that many phonologists
believe that the stream of speech can be divided into tone units within which pitch
moves. Pitch movement is expressed basically in terms of four categories: prominence,
tone, key, and termination (Brazil, 1985, cited in Coulthard, 1985: 101). The first, as
Brazil (1994b: 9) makes the point, is better regarded as something one can recognise
only within the overall pattern of the tone unit of which it is part . Thus, prominence can
only be considered a category of pitch movement in so far as it lends itself as the key
element upon which pitch moves. In addition, it seems prominence can be described, in
many cases, in the same way is stress in dictionary entries.
Compare:
Dictionary entry
Tone unit
a) /n
//noTAtion//
te
n/
b) /k m ju n ke
n/
//coMMUniCAtion//
Apparently, there is only a shift in terminology in the comparison above. For one thing,
when a one-prominent-syllable word uttered in citation form, as in a), is spoken as a
tone unit, its stressed syllable is labelled tonic syllable (notice that, following Brazil s
(1994a, b) suggestion, tonic syllables are capitalized and underlined to distinguish them
from other prominent syllables, which are also capitalized). For another, when a word
uttered in citation form, as in b), has both a secondary and a primary stress, the syllable
on which the former is laid is coined prominent syllable, and the one where the latter is
placed is the tonic syllable when it is spoken as a tone unit (Brazil, 1994b: 10;
McCarthy, 1991: 99). However, the very fact that the tonic syllable is always the last
prominent syllable in a tone unit seems to prevent further analogy between dictionary
entries and tone units in terms of prominence.
Compare:
Dictionary entry
Tone unit
14
/
i z pe r
/
either //CHEESEparing// or
//cheesePARing//, or yet //CHEESEPARing//
but not //CHEESEPARing//
Thus, it is evident that, in the case of words having the same pattern as cheeseparing
when spoken in citation form, the second prominent syllable either becomes nonprominent when spoken as a tone unit since normally there is no prominence after the
tonic syllable within a tone unit (Brazil, 1994b: 34), or becomes the tonic syllable of the
tone unit, obliterating then the similarity between dictionary entry and tone unit stress
pattern. McCarthy seems to concur with that:
Many other polysyllabic words may only have one prominence but may still have
primary and secondary word stress (e.g. 1CAta2lyst, 1CONfis2cate,
1
WHEREa2bouts).
(McCarthy, 1991: 95)
The distinction between dictionary-entry and tone-unit stress pattern appears to have
relevant implications for FPT. For one thing, if learners are denied this notion, they may
have difficulty in both producing intelligible and non-stilted utterances and in
understanding native-speaker production. This is very much the case when the
misleading belief that dictionary-entry phonemic notations provide the only correct
pattern for pronunciation is part of learner conception of language learning. For another
thing, even with those mature learners trained to notice the difference between these two
systems of stress pattern there is some risk of evasiveness of meaning if they lack the
awareness of the significance of prominence.
2.3.3 The significance of prominence: A speaker s decision of giving prominence to certain
words seems to be dependent on the available choices in the existential paradigm, the set
of options available in a given context. This is amply exemplified in a set of three
question/response pairs found in Coulthard (1985: 102), which is quoted here:
1. Q: Which card did you play?
R: //the QUEEN of HEARTS//
2. Q: Which queen did you play?
15
R: //the queen of HEARTS//
3. Q: Which heart did you play?
R: //the QUEEN of hearts//
Considering the composition of the pack of cards, as Coulthard (1985:103) suggests, in
(1) the speaker had a universe of thirteen possibilities to choose from on the one hand
and another of four possibilities on the other. However, in (2) the number of
possibilities drops to four, since queen is a given term, that is, queen is part of the
shared knowledge between the speakers. In the third example, there is a universe of 13
possibilities, inasmuch as the newly shared information is no longer focused on the
suits. In sum, the relevant aspect about prominence is that a speaker will make
prominent (1) items that have an existential paradigm to draw on (2) on the basis of
what he/she considers as not part of the listener s common ground, or (to use
Coulthard s terms) not interactionally given (Coulthard, 1985: 104).
There is a moot point over the characterisation of tone, the second category, labelled as
pitch movement by many discourse analysts (Coulthard, 1985:102; Sinclair and Brazil,
1982: 102; Brazil, 1994b:8). It seems though that Brazil s (1994a, b) simplified
terminology is appropriate for work at suprasegmentals in the L2 classroom (Table 2).
Additionally, although there is also broad disagreement amongst linguists as to the exact
meaning pitch movement conveys, the relevance of tone variation seems to be precisely
that it may have a different meaning whenever it occurs. The prime example of this is
the choice speakers seem to make between, say, a falling tone and a fall-rise tone
proclaiming, symbol p, and referring, symbol r, respectively, to use Brazil s
terminology (Brazil, 1985, cited in Coulthard 1985: 106, and Cauldwell and Allan,
1998: 30-31).
Compare:
a) //r he ll be TWENTY//p in AUgust//
b) //p he ll be TWENTY//r in AUgust//
16
With these example, Coulthard (1985: 106) demonstrates Brazil s (1985) argument that
referring tones are chosen when the speaker wants to convey parts of his/her message as
part of the shared knowledge with the listener, and that proclaiming tones are used as an
indication of addition of new items to the area of shared knowledge. Thus, in example
a) a potential hearer is told
when a mutual acquaintance will have his twentieth
birthday , whereas in b), the suggestion is that the potential hearer is told how old the
acquaintance is (or will be) . There is therefore an obvious correspondence between
interlocutors shared knowledge and referring tones on the one hand, and what is news
and proclaiming tones on the other.
Importantly, the notion of common ground between interlocutors can be at the speaker s
service for the purpose of ideological manipulation. Underhill (1994:86) points out that
this is very much the case in the speech of advertisers and politicians who may use
intonation to suggest that what they are saying is already negotiated and agreed by us,
and part of our common ground, even when it isn t .
Moreover, Brazil (1994b: 20; 60) also adds some equally important facts about
proclaiming and fall-rise tones. He points out that while r is the preferred tone in (a)
making-sure questions and in (b) social enquiries, p is normally the choice in (c)
enquiries about matters unknown to the enquirer and (d) when information asked for is
provided.
Compare*
a) //r i suPPOSE you don t know who the PUBblisher is// (a shop assistant to a
customer in a bookshop)
b) //r are you enJOYing england// (a member of a host family to an exchange student)
c) //p WHERE S the nearest Telephone please// (an old woman to a police officer)
*Utterances a and b are quoted from Brazil (1994a: 42-43; 45)
d) //p it s next to the GREEN BUILding madam// (the officer s response to the
woman s enquiry)
17
Table 2
SYSTEM OF PITCH MOVEMENT
Tone
Symbol
Falling
p
Rising
r+
Rise-fall
p+
Fall-rise
r
Level
o
Shape
The choice between fall-rise (r) and rising (r+) tones equally amply exemplifies the fact
that a slight shift in pitch may represent a shift in meaning. Nonetheless, although
Sinclair and Brazil (1982: 114-119) seem to have chosen not to explore the
subconscious ideological edge involved in the r/r+ system, the choice of either one is
likely to be related with the role of the speakers involved (Sinclair and Brazil, 1982:
116-119). For one thing, as Sinclair and Brazil (1982) suggest, the choice of r+ is likely
to be made by members of a specific class of interlocutors, such as teachers,
interviewers, employers, and senior officers/employees, or by anyone who arrogates to
themselves the role of dominant speaker. This seems to occur in extremely restricted
environments, and occurs mostly because of the strength of long-established social rules
as, for instance, it is the case of teacher/pupil interaction:
This is not to say that pupils can t [reciprocate what teachers do], it is just that it is
unusual and if they do so they are seen as doing something different from the
teacher, usually they are being cheeky .
Coulthard, 1985: 109
18
Brazil (1985, cited in Coulthard, 1985: 110) argues that the option realised by p+ in
detriment of p is used to indicate addition of simultaneous information both to the
common ground and to [the speaker s] own store of knowledge . Furthermore, he goes
on to say that p+ is also used to express feelings, such as surprise, disappointment, and
enjoyment.
e.g.
//p+ you ve MISSed the PLANE//
On the face of it, the decision on whether p+ signals surprise, disappointment or
enjoyment is dependent not so much on the pitch movement proper as on context (facial
gestures inclusive), as it seems to be the case in the aforementioned example. At this
point, many would be likely to agree that discourse analysts would capitalise much
more on Crystal s (1995) comprehensive description of tone forms than on Brazil s
simplified description of pitch movement.
Sinclair and Brazil (1982: 41) point out that key, the third category, has to do with the
pitch level of the voice , and that it describes spoken interaction in three levels, each
conveying a different meaning (Table 3). The examples in Coulthard (1985: 111)
quoted in Table 3 seem to lend credence to that.
Table 3
THE KEY SYSTEM
Pitch level
Meaning
Example
//p he GAMbled// p and LOST//
(contrary to expectations; i.e.
High key
Contrastive
there is an interaction-bound
opposition between the two)
//p he GAMbled// p and LOST//
Mid key
Additive
(he did both)
//p he GAMbled// p and LOST//
(as you would expect, i.e.
Low key
Equative
there is an
interaction-bound equivalence
between them)
In the light of these examples, it seems safe to suggest that pitch levels are items in the
intonation system which speakers can use to convey meaning at any given level of oral
communication. Additionally, apart from the three meanings described in Table 3, key
19
can be used to express attitudes such as politeness, assertiveness and indifference
(Coulthard, 1985).
Finally, there is termination. Sinclair and Brazil (1982:152-154) demonstrate that
termination is intrinsically related to key, and that the number of termination choices is
limited (Table 4). The examples in Table 4 are adapted from Sinclair and Brazil
(1982:154-155).
At least two assumptions may be made from the composition of Table 4: one is that
termination can move only one step up or one step down from de chosen key level
(Sinclair and Brazil, 1982: 154); and the other is the fact that dominant speakers seem to
capitalise on termination level (Coulthard, 1985: 118).
Sinclair and Brazil (1982) maintain that discourse analysts can choose to focus on a few
termination choices, namely those involving the transition of turns in an interaction.
Furthermore, they argue that in such instances termination is meaningful insofar as one
speaker s termination choice suggests his/her expectation about the termination choice
for the response when handing turns over to the other interlocutor. Additionally, it
seems that this is the case even when a turn is handed over with low termination, since
the speaker s expectation here is exactly not to expect his/her listener response in any
particular key-level choice.
Many would be likely to agree that the introduction of suprasegmentals in the L2
classroom, especially that of the intonation systems of prominence, tone, key and
termination, cannot be seen as an easy task
and it is certainly not for non-native
teachers because of the limitations of L2 linguistic experience, a phenomenon that is
part of FL learners reality (Harnsberger, 2001). The assumption in Brazil s (1994a, b)
Pronunciation for Advanced Learners of English is that awareness of these systems is
not only pivotal for the enhancement of oral/aural fluency, but also for the development
of skills in the field of semantics at an advanced level.
20
Table 4
THE TERMINATION SYSTEM
Key
Termination
level
level
High
Meaning
Example
Expectation on the part of the
T: //p do you THINK this one s
speaker of a contrastive
LARger
//
answer: yes/no.
P: //p YES//
High
Mid
Pressure from the speaker for
T: //p do you THINK this one s
a favourable response by the
LARGer//
other interlocutor.
P: //p YES//
High
Expectation on the part of the
speaker of a contrastive
answer: yes/no.
Mid
Mid
Pressure from the speaker for
a favourable response by the
other interlocutor.
Low
Mid
LARger
//
P: //p YES//
T: //p do you THINK this one s
LARGer//
P: //p YES//
The speaker imposes little or
//r and the QUEStion i want to
no constraint on the next turn
PUT to you//r+ IS//p DO we
taker .
NEED//p an INcome POLicy//
Pressure from the speaker for
a favourable response by the
other interlocutor.
Low
The speaker imposes little or
Low
T: //p do you THINK this one s
no constraint on the next turn
taker .
T: //p do you THINK this one s
LARGer//
P: //p YES//
//r and the QUEStion i want to PUT
to you//r+ IS//p DO we NEED//p an
INcome POLicy//
21
However, the great challenge now seems to adapt the wealth of information on DI
shared by the aforementioned authors to beginning-level L2 classrooms. Given the
importance of suprasegmental features for successful communication, it appears that the
earlier L2-learner consciousness about DI is raised the earlier pronunciation quality and
ability to express and perceive meaning are improved (Anderson-Hsieh et al., 1992;
Derwing et al., 1998, all cited in Rossiter and Derwing, 2002).
In the light of this brief review on the role of pronunciation in the L2 classroom in the
perspective of some language-oriented writers and teachers, it seems that (1) FPT is still
an option for those seeking good quality pronunciation; that (2) an eclectic approach
towards FPT appears to be more in line with modern views on it; (3) and lastly but
equally important is that FPT appears to be currently capitalising on suprasegmentals.
Nonetheless, there seems to be a crucial question left unanswered in the literature
reviewed here: Will an FPT approach involving the awareness of a given TL sound
system through phonemic symbols, at the segmental level, and DI, at the
suprasegmental level, be more effective in aiding beginning learners in the production
of intelligible TL utterances than one focusing solely on traditional pronunciation drills?
In the pursuit of the answer to this question, and with the intent to hopefully make the
slightest contribution to this field, the tentative micro experiment detailed in the next
chapter was implemented.
22
CHAPTER 3
THE EXPERIMENT
3.1
The research question
The research in this study is based on the following question: Will beginning learners
of English as a foreign language submitted to a treatment of explicit pronunciation
teaching, involving awareness of phonemic symbols (segmental phonology) and
consciousness-raising activities based on Brazil s (1994a, b and elsewhere) DI
(suprasegmental phonology) (T1), display more accuracy and intelligibility in their
utterances than those trained solely based on pronunciation drills, exposure to fluent
speakers on recordings and implicit acquisition of pronunciation features (T2)?
3.2
The research hypotheses
At first, it was hypothesised that (1) beginning learners submitted to T1 would display
significantly higher accuracy and intelligibility pronunciation production levels than
those submitted to T2.
Nevertheless, in the absence of empirical evidence in the
literature to suggest that T1 is more effective (or otherwise) than T2, two other
hypotheses had to be formulated: (2) that learners receiving T2 would significantly
outperform those receiving T1; and that (3) there would be no significant difference in
learners pronunciation production quality when comparing the speeches of a group
submitted to T1 with those of another group submitted to T2. As an attempt at verifying
these hypotheses, an experiment was designed and implemented at a state school in
Brazil.
3.3
Subjects
The experiment was conducted on a sample of sixteen Brazilian students, six girls and
ten boys, in the 13-16 age range, attending the sixth year of Ensino Fundamental (EF) at
Inspetora Dulcinéia Varela Moura State School, in Novo Israel, a district in the west
zone of the city of Manaus, the capital of the state of Amazonas. The subjects were
23
divided into two eight-student groups: the control group (CG), and the experimental
group (EG). The criteria adopted for standardizing CG and EG participants were
essentially (1) the level of familiarisation displayed by candidates with the
pronunciation of the words involved in the experiment (obtained through a reading test),
(2) candidates rate (in hours) of monthly exposure to spoken English, (3) candidates
attitude towards learning English and (4) their level of motivation towards acquiring
native-like pronunciation of a target language. Additionally, the subjects names are
replaced with numbers in this report in compliance with the National Research Act of
1974 (cited in Seliger and Shohamy, 1989:196).
3.4
Material
The selection of CG and EG involved two strands: a questionnaire (Appendix I), which
was completed by two EF sixth-year groups, totalling forty-one students, and a reading
test consisting of a thirty-seven-word text (henceforth TX) and a list of thirty-three
entries (Appendix II). Furthermore, care was taken to select a text containing sounds
which are non-existent in the sound system of Brazilian Portuguese.
Importantly, the selection test administered to these candidates was considered the
pre-test of the participants in the experiment, and was used for post-testing them as well.
The reading of the word list and TX (the items constituting the pre- and post-tests) by
each CG/EG member was recorded on two TDK-60 microcassettes, with a GE 3-5373
microcassette recorder.
In addition, the choice of a word list and a text to constitute the pronunciation
production test (PPT) in the study was based on the suggestion that the quality of
students pronunciation is improved in more controlled situations (Dickerson, 1975;
Koren, 1995: 392). In Koren s rising degree of care in pronunciation
an adaptation of
Tarone s (1983: 152, cited in Koren, 1995) interlanguage continuum
the reading of a
word list is categorised as more attended speech data and that of a text as attended
speech data . A further benefit from this kind of PPT is that it seems to diminish oraltest assessment subjectivity.
24
For exposing CG and EG members to the NSP used in the experiment, a Panasonic RXD14 portable stereo CD system and the cassette two (henceforth PALE-KT) of Brazil s
(1994a, b) Pronunciation for Advanced Learner's of English (PALE) were used.
3.4.1 Worksheets used with the Control Group
The treatment applied to the control group involved two worksheets (Appendix III): a
list of words from TX (Worksheet A), and TX (Worksheet B); a translation of the items
into Portuguese was provided on both worksheets.
3.4.2 Worksheets used with the Experimental Group
Because of the nature of the treatment administered to the experimental group, five
worksheets had to be used with this group. For the sake of practicality, they will be
labelled in alphabetical order following the sequence of the ones used with the CG
students, thus from C to G.
Worksheet C consisted of a list of TX words coded in phonemic transcriptions
following Underhill s (1994:49) two-level patterns for pronunciation practice plus their
respective Standard English spelling forms
the phonemic symbols used in the research
can be found in the Cambridge International Dictionary of English (1995). Importantly,
at this point, EG had not been provided with a Portuguese translation of the English
words involved in the experiment yet; this was purposefully postponed as an attempt at
having EG members focus on sound image in the early stages of the treatment.
The next handout (worksheet D) presented TX divided into tone units, an adaptation of
Cauldwell and Allan s (1998: 46) analysis of a larger portion of the same excerpt found
in unit eight of the PALE Course (Brazil, 1994a, b). Worksheet E had all elements from
worksheet C, except for the level-1 column, plus the Portuguese translation of the words
involved. Worksheet F comprehended an exercise on phonemic transcriptions involving
Portuguese and English words.
25
Finally, worksheet G included two TX versions: the first consisted of the content of
worksheet B, and the second was TX divided into tone units. Additionally, for the
treatment administered to EG, a collection of thirty-two cards (size = 8.7 in. x 9.8 in.)
displaying triple information: (1) phonemic symbols, (2) pictures and (3) captions both
in Portuguese and English was devised (Appendix IV)
3.5
Procedure
After taking the pre-test, CG and EG were submitted to two different sequences of
activities aiming at verifying the research hypotheses. An estimated total of 90 minutes
was spent on the treatment administered to each group: 30-minute afternoon meetings
were held once a week. Furthermore, the treatment administered to CG consisted of two
simple activities: (1) reading through Worksheet A several times, (2) and trying to
mimic a native speaker of English on the PALE-KT excerpt used in the research
(Worksheet B)
this recording was the NSP. The conduct of these activities followed
traditional models of interaction, such as teacher
and teacher
whole class; teacher
small groups;
individual student.
In contrast with CG, EG treatment was far more challenging for the students in the
group. First the students were introduced to the thirty-two phonemes of the English
language employed in the construct of the speech sequence used in the experiment.
Symbols of the International Phonetic Alphabet (IPA) were used for labelling the
phonemes: nine monothongs [i ], [
diphthongs [e ], [a ], [a ] and [
], [ ], [ ], [e], [æ], [ ], [ ] and [ ]; four
]; and nineteen consonants [p], [t], [k], [f], [ ],
[s], [ ], [b], [d], [v], [ð], [z], [l], [r], [j], [w] [m], [n], and [ ].
The introduction of these speech sounds represented by IPA symbols seemed to be in
line with the visual and kinaesthetic approach towards pronunciation teaching. The
materialisation of sounds through graphic representations and notions on place and
manner of articulation of phonemes are reported to be useful resources in helping
enhance L2 learner pronunciation production quality (Acton, 1984; Gilbert, 1993;
26
Brazil, 1994a, b; Pennington, 1996). While achieving the latter, kinaesthetic sensation,
was attempted through training on individual sound production, the redundancy of the
triple information in the aforementioned cards (Appendix IV), which was meant to be a
mnemonic strategy to accelerate EG learning, considering the short time span of the
study, seems to have served the purpose of former.
The next step was an attempt at making EG aware of the coarticulation phenomenon;
this was done by employing the discovery technique suggested by Underhill (1994).
Working at worksheet C, EG students first uttered slowly and discretely the phonemes
forming the construct of the words in the list; they were then asked to merge these
phonemes; this aimed at having the students utter the words in a seamless flow
(Underhill, 1994: 49).
Worksheet D was a consciousness-raising device to give EG rudimentary notions on the
intonation systems of prominence, tone, and key as posited by Brazil (1994a, b). A
contrastive analysis approach was adopted at the onset of the activity involving this
handout. That is, first some utterances in Portuguese were analysed on the basis of
Brazil s intonation system, then worksheet D was introduced.
The third activity was a brief repetition session. This time EG, following worksheet E,
had access to the translation of the data into Portuguese. The next handout (worksheet
F) was assigned as homework. Finally, worksheet G was used in a fifteen-minute
mimicry session where EG listened to NSP on the PALE-KT and tried to imitate it. The
post-test was then administered for posterior analysis in conjunction with the pre-test.
27
CHAPTER 4
THE ANALYSIS OF THE QUESTIONNAIRE, THE PRE- AND POST-TESTS
4.1
The questionnaire
The questionnaire, initially administered to the forty-one candidates from which CG and
EG were selected, was designed to elicit three pieces of information. The first was the
monthly amount of time that the students were exposed to spoken English through seven
different media, namely music (M), the cinema (C), language schools (LS), state schools
(SS), audio/video recordings with the purpose of learning English (LSP), radio (R) and
television (TV) programmes (Figure 1; see also Table 1 in Appendix V). The second
was the students attitude towards learning English. This was attempted through a set of
very simple questions (Figure 2; see also Table 2 in Appendix V). The third elicitation
was the importance the students attached to achieving native-like pronunciation of a
given target language and their disposition to participate in the experiment (Figure 3; see
also Table 3 in Appendix V).
By having the candidates answer the first part of the questionnaire, the objective was to
choose students having as little exposure to spoken English as possible. Thus, candidates
having less that 10.3-hour (or a maximum of 22 per cent of the total hours of exposure
of all candidates) monthly exposure participated in the experiment (Table 1, in
Appendix V), since the rate range and the media through which these students were
exposed to spoken English did not represent significant variables in the study, as it was
confirmed afterwards by the pre-test results (Table 6).
The answers to the second part of the questionnaire revealed interesting facts about the
candidates attitudes and motivations towards learning English. For one thing, although
ninety-five per cent of these students expressed the desire to learn English, the need to
do so seems to be restricted to basically being able to be successful in university
entrance and/or proficiency examinations (88%), and for increasing their chances in the
work market (85%).
28
Figure 1
CG and EG monthly exposure (in hours) to spoken English (the random numbers are
replacing the participants' names out a less-to-more-time-of-exposure rank list of 41
students)
9
8
7
Time line
6
5
4
3
2
1
0
2 6 18 13 21 25 28 30 38 40 5 12 14 22 34 17
Candidates selected for the experiment
M
C
LS
SS
LSP
R
TV
For another thing, thirty-seven per cent of the respondents said they need English to be
able to understand songs and films; but only a few of them (17%) expressed the
necessity to learn English to read publications (Figure 2; see also Table 2 in Appendix
V).
Perhaps the reason for the respondents little interest in learning English for reading
publications demonstrated in Figure 2 is corroborative evidence that the celebrated
crise da leitura
(reading crisis) in Brazil is generated much more by poor
socioeconomic conditions and the inefficiency of the educational system of this country
than by the dearth of reading materials and readers , for instance (Martins, 1982: 27).
29
For estimating the value candidates attached to good pronunciation when learning an
L2, the students were asked whether pronunciation accuracy or success in
communication, regardless of faulty pronunciation, was top priority for them. Ninety per
cent of the students responded that they would rather display good pronunciation.
Similarly, when asked if they would like to achieve native-like pronunciation, eightyfive per cent of the respondents answered favourably.
Figure 2
Proportion of candidates giving y-answers
Rate of candidates' attitudes towards learning English
100
90
80
70
60
50
40
30
20
10
0
A
B1
B2
B3
B4
Elicitation items
Symbols and terms:
A = I need to learn English
Reasons for learning English:
B1 = to take university entrance / proficiency examinations
B2 = to understand songs / films
B3 = to read publications
B4 = for my future career
Paradoxically, however, only thirty-nine per cent of the respondents volunteered to
participate in the experiment on FPT (Figure 3; see also Table 3 in Appendix V).
30
Importantly, in compliance with the National Research Act of 1974 (Seliger and
Shohamy, 1989:196), candidates who gave an n-answer to C4 were not selected for the
experiment despite high tally of y-answers and /or little exposure to English. Apart from
that, candidates tallying a minimum of three y-answers in section two of the
questionnaire (Appendix I) were assigned to the experimental and control groups.
Figure 3
Respondents' attitudes towards pronunciation accuracy
Proportion of answers
100
80
60
40
20
0
C1
C2
C3
C4
Elicitation items
y-answers
n-answers
ø (no answer was given)
Symbols and terms:
C1 = In my view, pronunciation accuracy is top priority when learning an L2.
C2 = In my view, communicability, despite faulty pronunciation, is top priority when
learning an L2.
C3 = I would like to achieve native-like pronunciation of an L2.
C4 = I would volunteer to participate in an experiment on English pronunciation
teaching.
Ultimately, as far as the questionnaire is concerned, the choice of CG and EG
participants was made essentially on the basis of candidates answers to the first section
and their disposition to join the experiment, which is recorded in the third section.
Apparently, most of the respondents answers to other items of this section and to all
31
items of the second section would not compromise the research findings were the
participants in the experiment chosen based solely on them.
4.2
Independent raters assessment of the pre- and post-tests
Three raters assessed CG and EG performance on both the pre- and post-test: one
American teacher of English (R1), one English painter (R2), and one English-Guyanese
teacher of English (R3). However, while R1 assessed only the first part of the pre- and
post-tests (the reading of the list of words and phrases) and R2 only the second part of
both tests (the reading of TX), R3 assessed both parts of the tests. This is the reason
why there is one application of the t test for each part of the tests, as shown in Tables 6
and 8. Furthermore, these assessments were based on the scales in Table 4 (Appendix
V).
The estimate of the inter-rater reliability was calculated by the formula:
r=
n ( XY) ( X) ( Y)
[n X2
( X)2] [n Y2
( Y)2]
The calculation of this reliability indicates that only in two occasions the raters scorings
correlate highly. (1) r = .83 (in R1 and R3 s sets of scores for the first part of the test)
and (2) r = .72 (in R2 and R3 s sets of scores for the second part of the test) on the pretest administered to the control group (Table 6). Furthermore, only sixty-eight percent
of the variation found between R1 and R3 s scorings for the experimental group on the
first part of the pre-test was reliable (r = .68). Additionally, the correlation between R2 s
and R3 s scorings on the second part of this test was zero percent (r = .0) (Table 6).
Figure 4 shows that, except for CG scores on the reading of the word list (r = 0.89), the
correlation between the scores given by all three raters is very low for both groups on
the post-test.
32
Figure 4
Estimate of the correlation between inter-raters' scores on the posttest for the experimental and control groups
1,00
0,90
0,80
0,70
0,60
0,50
0,40
0,30
0,20
0,10
0,00
R1 vs. R3
FPPT (CG)
FPPT (EG)
R2 vs. R3
SPPT (CG)
SPPT (EG)
Symbols and terms:
FPPT (CG) = First part of the post-test administered to the control group
FPPT (EG) = First part of the post-test administered to the experimental group
SPPT (CG) = Second part of the post-test administered to the control group
SPPT (EG) = Second part of the post-test administered to the experimental group
4.3
The statistical analysis of the tests
The statistical analysis of the pre- and post-test recordings by CG and EG members was
made according to conventional statistics criteria (Woods et al., 1986; Brown, 1988).
Because of lack of empirical evidence to support directionality, null hypotheses were
adopted. Moreover, the t statistic was applied to test the hypotheses (Table 5 in
Appendix V shows critical values of t, an adaptation from Fisher and Yates, 1963, cited
in Brown, 1988) and the alpha decision level was established at
< .01, nondirectional.
Brown s (1988) practical examples on the principle of dependency were most useful for
establishing the independent variable (CG or EG members) and the dependent variable
33
(participants pronunciation production quality, as reflected by their scores in the preand post-tests) contained in this study.
The levels of significance between the dependent- and independent-samples means were
estimated by this t-test formula:
X1
t=
X2
S2 + S2
N 1 N2
The statistic symbols and terms used in Tables 6-8 and 10:
X1 = Mean for the control group
X2 = Mean for the experimental group
S1 = Standard deviation for the control group
S2 = Standard deviation for the experimental group
N1 = Number of participants in the control group
N2 = Number of participants in the experimental group
df = Degrees of freedom
tobs = The observed statistic
tcrit = The critical value for the observed statistic.
4.3.1 The pre-test: Besides establishing the initial level of the participants, in terms of their
familiarisation with the pronunciation of the TL words involved in the experiment, the
scores of the pre-test were used for comparison with those of the post-test as an attempt
to identify any possible score improvement within the same sample. Table 6 shows the
scores of CG and EG members per rater.
34
Table 6
Assessment of the pre-test administered to CG and EG members: reading of the
list of words and phrases
R1
CG
member
CG
R3
EG
Mean
member
score
EG
Mean
score
CG
member
CG
EG
Mean
member
score
EG
Mean
score
12
2.1
2
20
12
43
2
26
13
18.2
5
32.4
13
45
5
61
14
45.5
6
24.8
14
73
6
54
17
41.2
22
23.6
17
63
22
48
18
25.8
25
28.8
18
57
25
52
21
20.6
34
33.9
21
60
34
59
28
12.7
38
31.8
28
59
38
41
30
19.4
40
28.2
30
52
40
52
R1
R3
X1 = 23.19
X1 = 56.50
X2 = 27.94
X2 = 49.13
S1 = 14.29
S1 = 9.77
S2 = 4.82
S2 = 11.22
N1 = 8
N1 = 8
N2 = 8
N2 = 8
df = 14
df = 14
tobs = 0.891
tobs = 1.401
tcrit = 2.977 (p < .01)
tcrit = 2.977 (p < .01)
Assessment of the pre-test administered to CG and EG members: reading of TX
R2
CG
member
CG
Mean
score
R3
EG
member
EG
Mean
score
CG
member
CG
Mean
score
EG
member
EG
Mean
score
12
7.5
2
30
12
7.5
2
14.5
13
12.5
5
25
13
12.5
5
18
35
14
17.5
6
27.5
14
15.5
6
33
17
17.5
22
20
17
15
22
36
18
17.5
25
20
18
15
25
24
21
10
34
22.5
21
15
34
33.5
28
12.5
38
20
28
15
38
17
30
15
40
15
30
15
40
17.5
R2
R3
X1 = 13.75
X1 = 13.81
X2 = 22.50
X2 = 24.19
S1 = 3.78
S1 = 2.71
S2 = 4.82
S2 = 8.72
N1 = 8
N1 = 8
N2 = 8
N2 = 8
df = 14
df = 14
tobs = 4.032
tobs = 3.214
tcrit = 2.977 (p < .01)
tcrit = 2.977 (p < .01)
Although there is a marked difference between the scoring of R1 and R3 in the first part
of the pre-test, the results per rater strongly indicate that the difference between the
means for CG and EG occurred by chance alone (tobs < tcrit in both R1 and R3). In other
words, in the initial stage, both groups were at the same level regarding the
pronunciation of the English words used in the experiment. Nevertheless, R2 and R3 s
assessment of the second part of the pre-test lead to the conclusion that EG members
displayed some advantage over CG members in terms of TL pronunciation both at the
segmental and suprasegmental levels (tobs > tcrit in both R2 and R3 s assessment).
Despite the consistency between sample means and t test results, R2 and R3 s
assessment of this part of the test is inconsistent with that of the first part.
4.3.2 The post-test: The comparison between the independent-samples means calculated
from the post-test was fundamental for testing two hypotheses: (1) whether any
difference in the means for CG and EG had occurred by chance alone (Ho), or if (2) it
36
was due to the treatments the groups received (H1). To this end, the t statistic was used
(Table 7).
The results indicate that Ho had to be accepted, since in both parts of the test tobs < tcrit
(Table7).
Table 7
Assessment of CG and EG post-tests: reading of the list of words and phrases
R1
CG
Students
CG
Mean
score
R3
EG
Students
EG
Mean
score
CG
Students
CG
Mean
score
EG
Students
EG
Mean
score
12
7
2
23
12
43
2
38
13
26,4
5
35,8
13
48
5
65
14
63,3
6
31,2
14
83
6
61
17
54,2
22
26,4
17
73
22
59
18
38,8
25
32,4
18
58
25
59
21
39,4
34
41,5
21
72
34
56
28
12,7
38
34,5
28
55
38
51
30
24,8
40
29,7
30
45
40
52
R1
R3
X1 = 33.33
X1 = 59.63
X2 = 31.81
X2 = 55.13
S1 = 19.42
S1 = 14.77
S2 = 5.72
S2 = 8.31
N1 = 8
N1 = 8
N2 = 8
N2 = 8
df = 14
df = 14
tobs = 0.212
tobs = 0.75
tcrit = 2.977 (p < .01)
tcrit = 2.977 (p < .01)
37
Assessment of the post-test administered to CG and EG members: reading of TX
R2
CG
Students
CG
Mean
score
R3
EG
Students
12
52,2
2
13
47,5
5
14
67,5
6
17
52,5
22
18
47,5
25
21
50
34
28
50
38
30
50
40
EG
Mean
score
CG
Students
60
57,5
57,5
62,5
50
65
45
45
CG
Mean
score
EG
Students
EG
Mean
score
12
28
2
27
13
21
5
34,5
14
54,5
6
24
17
44,5
22
24,5
18
37
25
22,5
21
52,5
34
35
28
36
38
25,5
30
29
40
23,5
R2
R3
X1 = 52.15
X1 = 37.81
X2 = 55.31
X2 = 27.06
S1 = 6.47
S1 = 11.95
S2 = 7.73
S2 = 4.93
N1 = 8
N1 = 8
N2 = 8
N2 = 8
df = 14
df = 14
tobs = 0.888
tobs = 2.352
tcrit = 2.977 (p < .01)
tcrit = 2.977 (p < .01)
4.3.3 A within-groups comparison of the sample mean scores:
For the within-groups comparison of the students scores, an important factor that has
been taken into consideration is the dependence of the groups, because the comparison
now was between the pre- and post-test means within the same group, which was
submitted to a particular treatment. Thus, as one of the assumptions to the application of
the basic t test is that of independence between the groups involved, the students tests
were divided randomly into two groups within each sample, as suggested by Brown
38
(1988:165). One disadvantage of this, in this specific study, was that the samples were
reduced to four students in each group (Table 8)
Table 8
Comparison of within-groups means: pre- and post-tests (reading of the list of words and
phrases)
1
CG
EG1
S
Ta MS
S
TbMS
S
TaMS
S
TbMS
12
22.6
18
48.4
2
23
25
45.7
13
31.6
21
55.7
5
46.7
34
48.8
14
59.3
28
33.9
6
39.4
38
42.8
17
52.1
30
34.9
22
35.8
40
40.9
X1 = 41.4
X1 = 36.23
X2 = 43.2
X2 = 44.6
S1 = 17.17
S1 = 9.91
S2 = 10.62
S2 = 3.45
N1 = 4
N1 = 4
N2 = 4
N2 = 4
df = 6
df = 6
tobs = 0.1784
tobs = 1.6
tcrit = 3.707 (p < .01)
tcrit = 3.707(p < .01)
Comparison of within-groups means: pre- and post-tests (reading of TX)
CG2
EG2
S
Ta MS
S
TbMS
S
TaMS
S
TbMS
12
7.5
18
42.3
2
22.3
25
36.3
13
12.5
21
51.3
5
21.5
34
50
14
16.5
28
43
6
30.3
38
35.3
17
16.3
30
39.5
22
28
40
34.3
X1 = 13.20
X1 = 25.53
X2 = 44.0
X2 = 39.0
S1 = 4.22
S1 = 4.30
S2 = 5.08
S2 = 7.40
N1 = 4
N1 = 4
N2 = 4
N2 = 4
39
df = 6
df = 6
tobs = 9.34
tobs = 3.15
tcrit = 3.707 (p < .01)
tcrit = 3.707(p < .01)
Symbols and terms:
S = Student
TaMS = Pre-test mean score
TbMS = Post-test mean score
1
= Means of scores given by R1 and R3
2
= Means of scores given by R2 and R3
The objective of applying the t test here was to verify whether (H1) or not (Ho) there was
any significant improvement in the students scores after the treatment administered to
each group.
In the reading of words and phrases, Ho had to be accepted for both groups, since tobs <
tcrit (CG: 0.1784 < 3.707; EG: 1.6 < 3.707). However, the observed statistic was much
greater than the critical one for the control group in the reading of TX: tobs > tcrit (CG:
9.34 > 3.707; EG: 3.15 < 3.707). In other words, although H1 had to be consistently
rejected for the experimental group, it had to be accepted in the comparison between the
within-group means of the second part of the test for the control group.
4.4
Comparing the NSP systems of intonation with those of CG and EG members
Apart from the assessment made by the independent raters, it was deemed necessary to
verify the frequency with which DI instances (as described by David Brazil) appeared in
the speeches of the participants. This was made by the researcher himself.
The assessment consisted of a comparison, at the suprasegmental level, between the
speech of the native-speaker parameter used in the experiment and that of the groups
involved. To this end, the researcher used the adaptation of Cauldwell and Allan s
40
(1998: 46) analysis of the NSP speech sequence as a parameter (Worksheet D, in
Appendix III).
From the NSP speech sequence on the PALE-KT, it was possible to tally the following
instances of the intonation systems of tone, prominence and key, as established by the
DI approach (Brazil, 1994a, b): 12 tone units (TU); 5 onset prominences (OP); 12 tonic
syllables (TS); 8 falling tones (FT); 1 rising tone (RT); 1 fall-rise tone (FRT) 2 level
tones (LT) and 4 high keys (HK) (Table 9).
The objective of this analysis was to determine the frequency with which these instances
would occur (exactly as they occur in the NSP speech) in the speeches of CG and EG
members when reading TX in the post-test. Additionally, it was at first hypothesised
that, ultimately, EG would outperform CG because of the notional activities the EG
members were engaged in during the treatment that they were given. But a directional
analysis could not be used here, since there seems to be no evidence in the SLA
literature for that. In reality, considering the age range of the participants, this
hypothesis is opposed by Brown s (1992) suggestion that young learners may achieve
better results with imitation activities than with analytical ones. However, if the claim
that benefits of imitation drills are likely to be dependent on learner s aptitude for oral
mimicry (Kenworthy, 1987), then a balance can be struck here.
In any case, at least two hypotheses had to be tested. The first (Ho), being that there
would be no significant difference in frequency between the two groups vis-à-vis the
NSP DI instances; and the second (H1), that either EG or CG speeches would display
significantly higher frequency of the DI instances found in the NSP speech.
The verification of these hypotheses was obtained through the following steps. First
NSP was multiplied by 8, as shown in Table 9
the frequencies of each category in the
speech of the imaginary NSPs being exactly the same. These frequencies were then
compared with the frequency with which the NSP categories appeared in the speeches
of CG and EG members (Table 10).
41
Table 9
The frequency of DI instances in the speech of 8 imaginary NSPs
Categories
NSP
Row
TU
OP
TS
FT
RT
FRT
LT
HK
total
1
12
5
12
8
1
1
2
4
44
2
12
5
12
8
1
1
2
4
44
3
12
5
12
8
1
1
2
4
44
4
12
5
12
8
1
1
2
4
44
5
12
5
12
8
1
1
2
4
44
6
12
5
12
8
1
1
2
4
44
7
12
5
12
8
1
1
2
4
44
8
12
5
12
8
1
1
2
4
44
96
40
96
64
8
8
16
32
360
Column
total
Symbols and terms:
TU = Tone unit
OP = Onset prominence
TS = Tonic syllable
FT = Falling tone
RT = Rising tone
FRT = Fall-rise tone
LT = Level tone
HK = High key
Table 10 shows that, despite the substantial difference in the frequency with which
some categories appear in the speech of some of the subjects, the difference between the
total frequency of the NSP DI instances appearing in the speech of each sample was
very small, thus insignificant (283 instances in the EG speech against 285 in the CG
speech). This assertion is supported by the t test.
To use the t statistic with this new set of data, two adaptations were made to its syntax:
(1) the row totals of the frequency of each category were regarded as total scores (as in
a discrete-point test) and (2) the frequency of the NSP categories the maximum points
CG and EG members could score (Tables 9 and 10). The sample means were X1 =
35.63 and X2 = 35.38 (CG and EG) respectively. S1 = 5.63 and S2 = 6.93 were the
estimated sample standard deviations; and the observed statistic was evaluated as tobs =
42
0.079 with 14 df. The decision level established at
< .01 confirmed that there was no
significant difference between the total frequency of the NSP categories appearing in
the speeches of CG and EG members, since the observed statistic was less than the
critical statistic (0.079 < 2.977), as can be verified in Table 5 (Appendix V).
43
Table 10
Frequencies of NSP DI instances in the speeches of CG and EG members
CG
Categories
Row
EG
Categories
Row
members TU OP TS FT RT FRT LT HK total members TU OP TS FT RT FRT LT HK total
12
8
5
12
7
1
0
2
4
39
2
11
5
11
1
0
0
2
3
33
13
6
5
11
2
0
0
2
1
27
5
11
5
10
8
1
1
2
4
42
14
11
5
12
8
1
0
2
4
43
6
12
5
11
8
1
1
2
3
43
17
9
4
11
7
0
0
2
4
37
22
12
5
11
8
1
1
2
3
43
18
10
4
12
1
0
0
2
0
29
25
7
4
11
4
1
0
2
1
30
21
9
5
11
6
0
0
2
3
36
34
8
4
9
0
0
0
2
1
24
28
10
5
11
8
1
0
2
4
41
38
5
3
12
7
0
0
2
3
32
30
6
5
10
7
0
0
2
3
33
40
8
5
11
7
1
0
2
2
36
69
38
90
46
3
0
16
23
285
74
36
86
43
5
3
16
20
283
Column
total
Column
total
X1 = 35.63
X2 = 35.38
S1 = 5.63
df = 14
tobs = 0.079
tcrit = 2.977 (p<.01)
N1 = 8
Symbols and terms:
S2 = 6.93
N2 = 8
TU = Tone unit
OP = Onset prominence
TS = Tonic syllable
FT = Falling tone
RT = Rising tone
FRT = Fall-rise tone
LT = Level tone
HK = High key
44
CHAPTER 5
THE RESULTS REVISITED AND SOME CONCLUDING REMARKS
Ultimately, the statistical analysis of the mean scores achieved by the experimental and
control groups in the post-test should serve the purpose of testing three hypotheses:
Ho: There is no significant difference between the means for the EG and CG
members.
H1: The mean for the EG members is significantly higher than that for the CG
members.
H2: The mean for the EG members is significantly lower than that for the CG
members.
Clearly, the t statistic obtained from the post-test indicated that the null hypothesis had
to be accepted. That is, there was no significant difference between the means for the
experimental and the control groups. Apparently, this means to say that both an
approach towards FPT based solely on imitation of parameters of TL speakers or on the
awareness of the TL sound and intonation systems with no emphasis on repetition drills
will yield similar results in the attempt of enhancing the level of accuracy and
intelligibility of L2 beginning learners pronunciation production. However, some
factors seem to dismiss the parallel between these approaches based on the findings of
this study as inconclusive. Some of these considerations relate to the conditions under
which the experiment was conducted and others to the assessment of the pre- and posttests.
5.1
Factors related to the context of the investigation
Added to the fact that this tentative investigation was undertaken by a novice teacherresearcher, the conditions under which the experiment was conducted posed some
important variables. Firstly, because the sessions were held in the afternoon and the
classrooms were not equipped with air-conditioning system, the rooms were extremely
45
hot (+ 36°C on average). Such discomforting temperature in the classroom seems to
have prevented the students (and the teacher) focusing continuously on the lesson.
Secondly, the gap between sessions appears to have produced a negative effect on the
experimental group, precisely because of the consciousness-raising activities in which
they were engaged. After the first session, precious time had to be spent on
recapitulation before introducing new activities in the two subsequent meetings. Lastly,
and perhaps most importantly, a crucial variable seems to be the type of clientele taking
part in the experiment.
It is a well-known fact (at least it is the case in Brazil) that, in general, state-school
students show little motivation towards learning a foreign language. The reason for this
may have to do with a variety of factors ranging from student personal goals to teacher
qualification (Brasil, 1999: 147). The fact is that the students who had to do most of the
thinking (the EG members, since the activities assigned to the CG members were
essentially mechanical) during the experiment sessions were the least interested in
undergoing the treatment. Furthermore, students idle chit-chatting in class, lack of
punctuality, and absence (only 62.5% of the students in each group attended all the
sessions) were some of the elements evincing the fact that, despite their de dicto yanswers elicited with the questionnaire, the participants had not agreed de facto to
embark on the experiment.
Perhaps, the reproduction of this experiment in a language school, where apparently the
clientele is more motivated to learning the foreign target language, would yield more
conclusive results. But it is somehow difficult to implement such an experiment in an
institution where students pay for their schooling, and this was one of the reasons why a
state school was chosen in the first place.
5.2
Factors related to the assessment of the pre- and post-tests
The independent raters assessment of the pre- and post-tests also posed crucial
variables to the study. Seemingly, these raters were given a clear task. That is, they were
asked to determine the degree of accuracy and intelligibility of the subjects speeches
46
recorded on an audiocassette based on a native-speaker parameter which, in this case,
was the raters themselves. To this end, they had to follow the scales in Table 4
(Appendix V). This has not prevented them, however, from producing low correlated
scores for the pre- and post-tests.
In the pre-test, EG displayed some advantage over CG in the reading of TX. The
difference between the means for the two groups based on R2 s assessment (X1 = 13.75;
X2 = 22.50) (Table 6) is significant at the one per cent level (tobs = 4.032 is greater than
tcrit = 2.977 at p < .01). Apparently, this is consistent with R3 s assessment of this part of
the same test (X1 = 13.81; X2 = 24.19, where tobs = 3.214 is greater than tcrit = 2.977 at p
< .01). Nonetheless, the great variance between the inter-rater correlation coefficients
seems to indicate that conclusions based on the raters assessment of the pre- and posttests should be treated with caution. This is clearly the case in the estimate of r from the
scores for the reading of TX in the pre-test. While R2 and R3 correlation coefficient is
relatively high for CG (r = 0.72), it drops drastically in the assessment of EG for the
same part of the test (r = 0.0).
The insignificance of the difference between the means for the experimental and the
control groups in the second part of the post-test (tobs < tcrit at p < .01) established by the
t statistic also seems to be compromised by the unreliability of scores across raters. The
correlation coefficients, estimated from this part of the test, are very low (r = 0.60 for
CG; r = 0.50 for EG) (Figure 4). This is further confirmed by the variation found
between R2 and R3 s sets of scores. R2 produced, on average, lower mean scores for
CG and higher for EG (X1 = 52.15; X2 = 55.31), whereas R3 produced otherwise (X1 =
37.81; X2 = 27.06) (Table 7). Moreover, the inconsistency of judgement of the raters is
also evident in the assessment of the list of words and phrases in both the pre-test (r =
0.83 for CG; r = 0.68 for EG), and the post-test (r = 0.0 for CG; r = 0.50 for EG).
In view of such discrepancy in raters assessments, one may wonder, for instance, if the
confirmation of the null hypothesis (tobs < tcrit at p < .01) in both parts of the post-test
administered to CG and EG was due to the fact that both treatments had similar effect on
47
the participants, or to inter-rater low correlation. This, in turn, elicits at least three other
questions: (1) was inter-rater low correlation a consequence of the raters lack of
expertise in this kind of assessment, (2) the interference of their accent, or (3) was the
quality of the recordings a major variable?
In hindsight, I claim that all three aspects had a bearing on the scores produced by the
raters. For one thing, apart from not being oral test specialists, these raters speak English
with noticeably different accents, diverging somehow from that of the NSP used in the
experiment. While the accent of this NSP was closer to that of the English painter, it was
very different from those of the Californian and Guyanese teachers of English. For
another thing, the fact that the CG and EG members spoke with no, or little pause during
the recording session, added to extraneous background noise on the tape, seem to have
impaired the raters judgement.
It is important to point out that, other things being equal, the inter-rater low correlation
in this study may evince the fact that estimating examinees language ability can be
challenging, even when standardized scales are used, because of the subjectivity of the
task (Bachman, 1990: 37-38). Thus, subjectivity may be another important variable
affecting the scores produced by R1, R2 and R3. Consequently, even considering that
the t observed is less than the t critical in both parts of the post-test for both groups, thus
confirming Ho, the conclusion that both treatments would have similar effects in
beginning learners of English would be an untenable one.
5.3
The NSP DI instances used in the speeches of the CG and EG members
The results of the comparison of the frequency with which the NSP DI instances
occurred in the speeches of the CG and EG members seem to indicate that, despite being
statistically insignificant, the observed difference in frequency is meaningful. Contrary
to what had been hypothesised, in retrospect, it seems that CG members should have
displayed more similarities with NSP in terms of intonation because of their longer
period of exposure to the NSP speech on the PALE-KT. An implication of this seems to
be the fact that, even achieving the same results by applying either approach, activities
48
which help learners develop their thinking process
(1981:13)
the fifth skill, according to Ur
appear to be more attractive, from a formative perspective, than mechanical
ones.
Another aspect that need be considered concerning the analysis of the intonation
systems of prominence, tone and key in the speeches of the participants is the fact that
the students had to read a text, which may account for the excess of level tones in their
speeches (Appendix VI). Brazil (1994b: 91) posits that, when reading aloud, the reader
may decide to focus exclusively on form in detriment of meaning. This kind of reading,
which normally lengthens tone units to the full capacity of the reader s ability to utter
words at a time, renders a great deal of unnaturally-used level tones. Koren echoes the
awareness of that by pointing out that pronunciation tests involving free speech are a
better arbiter of normal pronunciation by a second language learner, since it makes more
sense to assume that the learner uses the language for speaking rather than for reading
aloud (Koren, 1995: 388).
One final thought. Although a number of variables have prevented definite conclusions
on one direction or the other vis-à-vis the treatments administered to the control and
experimental groups, the investigation has taught this novice teacher-researcher various
important lessons; perhaps the most important of which are that, after painstaking work,
the researcher is left with many unanswered questions, and that the conclusions drawn
by him or her may not do justice to the results of the investigation. This seems to be well
expressed in these lines of Carroll s Alice in Wonderland (quoted in Seliger and
Shohamy, 1989: 243):
It seems very pretty, she said when she had finished it, but it s rather hard to
understand! (You see, she didn t like to confess even to herself, that she couldn t
make it out at all.) Somehow it seems to fill my head with ideas only I don t
exactly know what they are!
49
APPENDIX I
The questionnaire
1. Write an average of the monthly exposure (in minutes/hours) you have to spoken
English through these media.
a) Music (_______________)
e)
Audio/video recordings for studying
English (____________)
b) Cinema (______________)
f)
Radio (_____________)
c) Language school (___________)
g) Television (______________)
d) State school (______________)
2. Check the statements which express your motivation(s) towards learning English (if
you check the No box in a), don t check any other boxes in this section).
a) I need to learn English ( Yes;
No).
b) Reasons why I need to learn English
to take university entrance / proficiency examinations ( Yes;
to understand songs / films ( Yes;
to read publications ( Yes;
No).
for my future career ( Yes;
No).
No).
No).
3. How important is pronunciation to you? Check the yes or no boxes.
a) In my view, pronunciation accuracy is top priority when learning an L2
( Yes;
No).
b) In my view, communicability despite faulty pronunciation is top priority when
learning an L2 ( Yes;
No).
c) I would like to achieve native-like pronunciation ( Yes;
No).
d) I would volunteer to participate in an experiment on English pronunciation
teaching ( Yes;
No).
50
APPENDIX II
The reading test
(This test was administered to all 41 candidates, and for post-testing the EG and CG
members)
Among the speakers at today's National Transport Conference will be Mister Tom Williams.
Mister Williams, I understand that having been a keen, not to say fanatical, motorist for most
of your life, you're now having second thoughts?
(Brazil, 1994b: 114)
a
among
at
be
been
Conference
fanatical
for
having
I
keen
life
Mister
most
motorist
National
not
now
of
say
second
speakers
that
the
thoughts
to
today's
Tom Williams
Transport
understand
will
your
you're
51
APPENDIX III
Worksheet A
A = um, uma
AMONG = entre
AT = na, no
BE = ser, estar
BEEN = sido, estado
CONFERENCE = conferência
FANATICAL = fanático
FOR = por (período de tempo)
HAVING = tendo
I = eu
KEEN = entusiasmado
LIFE = vida
MISTER = senhor
MOST = a maior parte
MOTORIST = motorista
NATIONAL = nacional
NOT = não
NOW = agora
OF = de
SAY = dizer
SECOND = segundo
SPEAKERS = palestrantes
THAT = que
THE = o, a, os, as
THOUGHTS = idéias, pensamentos
TO = para
TODAY'S = de hoje
TOM = Tom
TRANSPORT = transporte
UNDERSTAND = entender, compreender
WILL = verbo auxiliar também usado para falar do futuro.
WILLIAMS = Williams
YOUR = seu, sua, seus, suas
YOU'RE = você está, você é
52
Worksheet B
Among the speakers at today's National Transport Conference will be Mister Tom Williams.
Mister Williams, I understand that having been a keen, not to say fanatical, motorist for most
of your life, you're now having second thoughts?
(Brazil, 1994b: 114)
Entre os palestrantes da Conferência Nacional de Transporte de hoje estará o Senhor Tom Williams.
Sr. Williams, eu percebo que, tendo sido um motorista entusiasmado, para não dizer fanático, a maior
parte de sua vida, o Sr. está mudando de idéia agora?
53
Worksheet C
Level 1
/
/
/
m
Level 2
/ /
/
/
Standard English spelling
A
m
AMONG
/
/æt/ / t/
AT
/
/bi /
BE
/b
n/
/b n/
BEEN
/k
n
f
/f
n
æ t
/æ t
/ /
/b
i
/f
t
/
r
k
/ /f
/h
æ v
/a
/
/k
i
/l
a
n
/
/f næt k l/
FANATICAL
/f / / / f /
FOR
/ hæv
HAVING
f
I
/ki n
KEEN
/la f/
LIFE
/
/ m st /
MISTER
/
/m
st/
MOST
/ m
t r st/
MOTORIST
/
t
t
r
n
l
t
t
s
t
/
/ næ n l/
NATIONAL
/
/n t/
NOT
/
/na /
NOW
/ v/ / v/
OF
/se /
SAY
/ seknd/
SECOND
/ spi k z/
SPEAKERS
/ðæt/
THAT
/ð /
THE
æ
/n
/n
a
/
v
/s
e
/s
e
k
n
/s
p
i
k
/ð
æ t
/ð
/
/ /
v
/
/
/
/
/
/a / / /
/m
d
/
z
/
/
t
s
/t
/
/t
d
e
/t
m
/
n
l
/
/
s
/
CONFERENCE
/
/m
r
/ k nf r ns/
/
s
/t
/
n/
/m
/n
s
n
d
/w
l
/
/w
l
j
/j
/ /j
/j
/
/
/
z
s
p
s
t
m
/
/
t
/
æ n
d
z
/
/
ts/
THOUGHTS
/t /
TO
/t de z/
TODAY'S
/t m /
TOM
/ tr
TRANSPORT
nsp t/
/ nd stænd/
UNDERSTAND
/w l/
WILL
/ w lj m z/
WILLIAMS
/j / / j /
YOUR
/j /
YOU'RE
54
Worksheet D
//
aMONG the SPEAkers //
//
at toDAY'S //
//
NAtional TRANSport conference //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
i underSTAND that //
//
HAving been a //
//
KEEN //
//
NOT to say //
//
fa NAtical motorist //
//
for MOST of your life //
//
you're now having
SECond
55
THOUGHTS //
Worksheet E
Standard
English
spelling
Pronunciation
m
/
/æt/ / t/
/bi /
/b n/
/ k nf r ns/
/f næt k l/
/f / / / f /
/ hæv
/
/a / / /
/ki n
/la f/
/ m st /
/m
st/
/ m
t r st/
/ næ n l/
/n t/
/na /
/ v/ / v/
/se /
/ seknd/
/ spi k z/
/ðæt/
/ð /
/
ts/
/t /
/t de z/
/t m /
/ tr
nsp t/
/ nd stænd/
/w l/
/ w lj mz/
/j / / j /
/j /
um, uma
entre
AMONG
na, no
AT
ser, estar
BE
sido, estado
BEEN
CONFERENCE Conferência
fanático
FANATICAL
por (período de tempo)
FOR
tendo
HAVING
eu
I
entusiasmado
KEEN
Vida
LIFE
senhor
MISTER
a maior parte
MOST
motorista
MOTORIST
nacional
NATIONAL
Não
NOT
agora
NOW
De
OF
dizer
SAY
segundo
SECOND
palestrantes
SPEAKERS
que
THAT
o, a, os, as
THE
idéias, pensamentos
THOUGHTS
para
TO
de hoje
TODAY'S
Tom
TOM
Transporte
TRANSPORT
UNDERSTAND entender, compreender
verbo auxiliar também usado para falar do futuro
WILL
Williams
WILLIAMS
seu, sua, seus, suas
YOUR
você está, você é
YOU'RE
A
/ /
/
Portuguese Translation
56
Worksheet F
Fill in the blanks below with words from these two boxes.
FUNK, MAU, NOTA, PRAIA, NAVIO, LEITO, ÁRVORE, CORDA, IOGA, SENTA, XÍCARA, PESA, MANGA,
TEIA, CANECA, DATA, LEI, MALOCA, MAR, NETA, VELA, UÍSQUE, FEITA, LETRA, BOTAS, TECLA,
CARTEIRA
TODAY, TRANSPORT, TRANSPORT, MOST, MOST, BE, SPEAKERS, FANATICAL, MOTORIST, HAVING,
THAT, THOUGHTS, MISTER, MISTER, WILLIAMS, BEEN, YOUR, CONFERENCE, NATIONAL, WILL,
AMONG, AMONG, KEEN, KEEN, SECOND, NOW, UNDERSTAND, LIFE, THAT, SAY, YOU'RE, NOT
Word Table
Phonemic symbols
(sounds)
/p/
Word in Portuguese
having a similar
sound
pata
Phonemic
transcription
(Portuguese)
[ pat ]
Word in English
having a similar
sound
Phonemic
transcription
(English)
/ tr
nsp t/
/b/
[ b t s]
/bi /
/f/
[ fe t ]
/f næt k l/
/v/
[ vel ]
não há som
equivalente
não há som
equivalente
/ /
/ð/
não há som
equivalente
não há som
equivalente
/s/
[ sent ]
/z/
[ pez ]
/t/
[ te ]
/ hæv
/
/
ts/
/ðæt/
/ m
t r st/
/s pi k z/
/m
st/
/d/
[ dat ]
/t de /
/m/
[m l k ]
/ m st /
/n/
[net ]
/bi n/
/r/
[mar] (no falar
caipira)
/ k nf r ns/
/l/
[le ]
/w l/
/k/
[k nek ]
/ /
[ ma g ]
/ki n/
/ /
[ ik r ]
/j/
[j g ]
/j /
/w/
[ wi ski]
/ w lj m z/
/
m
/
/ næ n l/
/ /
[ l tr ]
/ m st /
/e/
[ tekl ]
/ seknd/
[kah te r ]
não há som
equivalente
não há som
equivalente
/ nd stænd/
/ /
/æ/
/ /
não há som
equivalente
não há som
equivalente
/ðæt/
/
m
/
/ /
[n t ]
/n t/
/i /
[n vi j ]
/ki n/
/
/
[
/
/
[k
hv ri]
hd ]
/ tr
nsp t/
/j /
/e /
[le t ]
/a /
[ pra ]
/la f/
/a /
/ma /
não há som
equivalente
/na /
/
/
não há som
equivalente
57
/se /
/m
st/
Worksheet G
G1
Among the speakers at today's National Transport Conference will be Mister Tom
Williams. Mister Williams, I understand that having been a keen, not to say fanatical,
motorist for most of your life, you're now having second thoughts?
(Brazil, 1994b: 114)
Entre os palestrantes da Conferência Nacional de Transporte de hoje estará o Senhor Tom
Williams. Sr. Williams, eu percebo que, tendo sido um motorista entusiasmado, para não dizer
fanático a maior parte de sua vida, o Sr. está mudando de idéia agora?
G2
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
at toDAY'S //
NAtional TRANSport conference //
i underSTAND that //
KEEN //
NOT to say //
you're now having
58
fa NAtical motorist //
SECond
THOUGHTS //
APPENDIX IV
The flashcards
These flashcards were devised for teaching the CG and EG members the English
phonemic symbols used in the construct of the speech sequence on the PALE-KT.
59
60
61
APPENDIX V
Table 1
Monthly exposure (in hours) to spoken English in a sample of 41 candidates
for the experiment on FPT (figures in brackets are percentages; = data used in the analysis)
Student
M
C
LS
SS
LSP
R
TV
Row total
1,3
2
1
0
0
5,3
0
0
0
6,3
1,3
6
1
0
0
5,3
0
0
0
6,3
1,3
18
1
0
0
5,3
0
0
0
6,3
1,5
13
2
0
0
5,3
0
0
0
7,3
1,5
21
2
0
0
5,3
0
0
0
7,3
1,5
25
2
0
0
5,3
0
0
0
7,3
1,5
28
2
0
0
5,3
0
0
0
7,3
1,5
30
2
0
0
5,3
0
0
0
7,3
1,5
38
2
0
0
5,3
0
0
0
7,3
1,5
40
2
0
0
5,3
0
0
0
7,3
1,7
5
3
0
0
5,3
0
0
0
8,3
1,7
12
1
2
0
5,3
0
0
0
8,3
1,7
14
1
0
0
5,3
2
0
0
8,3
1,7
22
3
0
0
5,3
0
0
0
8,3
1,7
34
1
2
0
5,3
0
0
0
8,3
1,9
17
2
2
0
5,3
0
0
0
9,3
35
3
0
0
5,3
2
0
0
10,3 2,1
10
2
2
0
5,3
2
0
0
11,3 2,3
16
2
0
0
5,3
4
0
0
11,3 2,3
39
4
0
0
5,3
2
0
0
11,3 2,3
11
1
2
0
5,3
4
0
0
12,3 2,5
15
3
0
0
5,3
0
0
4
12,3 2,5
19
1
2
0
5,3
0
0
4
12,3 2,5
37
3
0
0
5,3
0
0
4
12,3 2,5
41
1
2
0
5,3
0
0
4
12,3 2,5
23
7,5
0
0
5,3
0
0
0
12,8 2,6
20
2
2
0
5,3
0
0
4
13,3 2,7
36
4
4
0
5,3
0
0
0
13,3 2,7
8
3
0
0
5,3
6
0
0
14,3 2,9
31
1
4
0
5,3
4
0
0
14,3 2,9
24
1
4
0
5,3
4
0
0
14,3 2,9
4
7,5
0
0
5,3
2
0
0
14,8 3,0
27
7,5
2
0
5,3
0
0
0
14,8 3,0
7
7,5
0
0
5,3
2
0
0
14,8 3,0
9
7,5
0
0
5,3
2
0
0
14,8 3,0
29
4
2
0
5,3
4
0
0
15,3 3,1
33
4
2
0
5,3
4
0
0
15,3 3,1
32
1
2
0
5,3
4
0
4
16,3 3,3
1
3
4
0
5,3
0
0
4
16,3 3,3
26
7,5
0
0
5,3
2
0
4
18,8 3,9
3
7,5
2
0
5,3
8
0
4
26,8 5,5
Column total 124 (25) 42 (8,6) 0 228,1 (47) 58 (12) 0 36 (7,4) 487,6 100
62
Table 2
Rate of candidates attitudes and motivations to learning English
Students
A B1 B2 B3 B4 row tally of y-answers
12
y y y y y
5
21
y y y y y
5
20
y y y y y
5
13
y y y n y
4
17
y y y n y
4
18
y y y n y
4
7
y y y n y
4
9
y y
n y y
4
11
y y y n y
4
23
y y y n y
4
26
y y y n y
4
31
y y
n y y
4
33
y y y n y
4
35
y y y n y
4
36
y y y n y
4
2
y y
n n y
3
5
y y
n n y
3
6
y y
n n y
3
14
y y y ø ø
3
22
y y
n n y
3
25
y y
n n y
3
28
y y
n n y
3
34
y y y n ø
3
38
y n n y y
3
40
y y
n n y
3
30
y y
n n y
3
4
y y
n n y
3
8
y y
n n y
3
10
y y
n n y
3
15
y n n y y
3
16
y y
n n y
3
19
y y
n n y
3
24
y y
n n y
3
27
y y
n n y
3
32
n n n n n
0
37
y y
n n y
3
39
y y
n n y
3
1
y y
n ø ø
2
3
y y
n n n
2
41
y n n n y
2
29
n n n n ø
0
Column tally of y-answers 39 36 15 7 35
132
63
Table 3
Students
12
21
20
13
17
18
7
9
11
23
26
31
33
35
36
2
5
6
14
22
25
28
34
38
40
30
4
8
10
15
16
19
24
27
32
37
39
1
3
41
29
Total of
y-answers
n-answers
Ø
Candidates attitude towards pronunciation accuracy
In my view,
In my view,
communicability
pronunciation
despite faulty
I would like to achieve
accuracy is top
pronunciation is native-like pronunciation
priority when
top priority when
learning an L2
learning an L2
y
n
y
y
n
y
n
y
n
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
n
y
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
n
Ø
n
Ø
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
n
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
y
n
y
Ø
Ø
Ø
y
n
y
y
n
y
y
n
n
37
2
2
2
38
1
35
4
2
64
I would join a
project on English
pronunciation
teaching
y
y
n
y
y
y
n
n
n
n
n
n
n
n
n
y
y
y
y
y
y
y
y
y
y
y
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
16
25
0
Table 4
Scale 1: assessing subject reading of words and phrases in a list.
Component: pronunciation at the segmental level (PSL).
%
Appraisal per token
100 90 Utterance virtually as a native speaker.
80 70
Utterance sufficiently correct, can be understood without difficulty.
60 50
Utterance can be understood with little difficulty.
40 30
Utterance poorly articulated so that comprehension is difficult.
20 10
Utterance so poor that comprehension is extremely difficult.
0
Unintelligible.
Scale 2: assessing subject reading of TX.
Component: pronunciation at the suprasegmental level (PSSL)
Virtually native-speaker stress-timing, rhythm, and placing of stress,
100 90 intonation patterns and range of pitch within sentence; natural linking
of phrases.
Stress-timing, rhythm, placing of stress, intonation, etc. sufficiently
80 70
native-like as to make comprehension easy and listening pleasurable.
Stress-timing, rhythm, placing of stress, intonation, etc. sufficiently
60 50
controlled.
Foreign speech patterns make the subject occasionally difficult to
40 30
understand.
20 10 Foreign speech patterns severely impede comprehension.
0
Not intelligible, through faulty stress and intonation.
Scale 3: assessing subject reading of TX.
Component: PSL
100 90 All individual sounds virtually as a native speaker.
80 70 Most individual sounds virtually as a native speaker.
60 50 All sounds sufficiently correct, can be understood without difficulty.
Some individual sounds poorly articulated so that comprehension is
40 30
sometimes difficult.
20 10 Individual sounds so poor that comprehension is often impossible.
0
Unintelligible.
(Adapted from UCLES, 1991:13)
65
Table 5 Critical Values of t (adapted from Fisher and Yates, 1963, cited in Brown, 1988: 168)
66
APPENDIX VI
Structure of the use of the intonation systems of prominence, tone and key by CG and
EG members
Note that this is not an exact transcription of their speech; the focus here is only on the
intonation systems of prominence, tone and key.
The Control Group
Student 12
//
aMONG the SPEAkers //
//
//
will BE //
MISTER //
MISter WILLiams //
HAving been a //
//
for MOST of your life //
//
//
//
HAving been a //
NOW //
//
to SAY //
you're now having
at toDAY'S //
fa NAtical //
SECond
MOtorist //
THOUGHTS //
NAtional TRANSport CONference //
i underSTAND //
KEEN //
for MOST //
Having //
NOT //
OF //
THAT //
TO //
YOUR //
SAY //
LIFE //
faNAtical //
YOU RE //
SECond THOUGHTS //
Student 14
//
aMONG the SPEAkers //
//
NOT //
THAT //
mister TOM WILLiams //
MISter WILLiams //
// MOtorist //
TOM WILLiams //
KEEN //
Student 13
//
aMONG the SPEAkers //
will BE //
NAtional TRANSport conference //
i underSTAND //
//
//
at toDAY'S //
at toDAY'S //
NAtional TRANSport conference //
will be mister TOM WILLiams //
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
i underSTAND that //
KEEN //
NOT to say //
you're now having
67
fa NAtical //
SECond
motoRIST //
THOUGHTS //
Student 17
//
aMONG //
//
//
THE //
SPEAkers //
NAtional TRANSport CONference //
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
at toDAY'S //
will be MISter //
i underSTAND //
KEEN //
THAT //
NOT to say //
you're now having
Student 18
//
aMONG the SPEAkers //
//
CONference //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
at toDAY'S //
TOM WILLiams //
fa NAtical motorist //
SECond
THOUGHTS //
NAtional // TRANSport //
will be mister TOM WILLiams //
i underSTAND that //
KEEN //
NOT to say //
fanatical //
you're now having //
SECond //
MOtorist //
THOUGHTS //
Student 21
//
aMONG the SPEAkers //
//
WILL //
//
BE //
at toDAY'S //
NAtional TRANSport CONference //
mister TOM WILLiams //
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
i underSTAND //
KEEN //
THAT //
NOT to say //
you're now having
fa NAtical //
SECond
MOtorist //
THOUGHTS //
Student 28
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of //
at toDAY'S //
NAtional TRANSport CONference //
i underSTAND //
KEEN //
your LIFE //
NOT to say //
THAT //
fa NAtical motorist //
you're now having
68
SECond
THOUGHTS //
Student 30
//
aMONG the SPEAkers //
//
will BE //
//
//
NAtional TRANSport //
CONference //
MISter TOM WILLiams //
MISter WILLiams //
HAving //
// MOtorist //
//
at toDAY'S //
BEEN a //
i underSTAND //
KEEN //
THAT //
NOT //
TO //
SAY //
faNAtical //
for MOST of your life //
you're now having
SECond
THOUGHTS //
The Experimental Group
Student 2
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST //
at toDAY'S //
NAtional TRANSport CONference //
i underSTAND that //
KEEN //
NOT to say //
of your LIFE //
faNAtical motorist //
you're now having
SECond
THOUGHTS //
Student 5
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
at toDAY'S //
NAtional TRANSport CONference //
i underSTAND that //
KEEN //
NOT to say //
you're now having
69
fa NAtical motorist //
SECond
THOUGHTS //
Student 6
//
aMONG the SPEAkers //
at toDAY'S //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
NAtional TRANSport CONference //
i underSTAND that //
KEEN //
NOT to say //
you're now having
fa NAtical motorist //
SECond
THOUGHTS //
Student 22
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a //
//
for MOST of your life //
at toDAY'S //
NAtional TRANSport CONference //
i underSTAND that //
KEEN //
NOT to say //
you're now having
fa NAtical motorist //
SECond THOUGHTS //
Student 25
//
aMONG the SPEAkers //
at toDAY'S //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving //
//
for MOST //
//
having SECond THOUGHTS //
i underSTAND //
BEEN a //
KEEN //
of your LIFE //
NAtional TRANSport CONference //
THAT //
NOT to say //
YOU RE //
fa NAtical //
MOtorist //
NOW //
Student 34
//
aMONG the SPEAkers //
//
will be mister TOM WILLiams //
//
MISter WILLiams //
//
HAving been a KEEN //
//
for MOST of your LIFE //
AT toDAY'S //
NAtional TRANSport conference //
i underSTAND that //
NOT to say //
fa NAtical motorist //
YOU RE now // having SECond THOUGHTS //
Student 38
70
//
aMONG the SPEAkers //
//
will be mister TOM //
//
MISter WILLiams //
//
HAving been //
//
for MOST of your //
at toDAY'S //
NAtional TRANSport //
CONference //
WILLiams //
i underSTAND that //
a KEEN //
LIFE //
NOT to say //
fa NAtical //
you're now HAving //
MOtorist //
SECond //
THOUGHTS //
NAtional TRANSport //
CONference //
Student 40
//
aMONG the SPEAkers //
//
WILL BE //
//
at toDAY'S //
mister TOM WILLiams //
MISter WILLiams //
//
NOT to say //
//
for MOST of your life //
i underSTAND //
that HAving been a //
fa NAtical motorist //
you're now having SECond THOUGHTS //
71
KEEN //
REFERENCES
Acton, W. (1984) Changing Fossilized Pronunciation . TESOL Quarterly 18(1): 69-83.
Bachman, L. F. (1990) Fundamental Considerations in Language Testing. Oxford: Oxford
University Press.
Baker, A. (1977) Ship or Sheep. Cambridge: Cambridge University Press.
Bloomfield, L. (1933) Language. New York: H. Holt, Rinehart and Winston.
Bowler, B. and Cunningham, S. (1991) Headway: Upper-intermediate Pronunciation.
Oxford: Oxford University press.
Brasil, Ministério da Educação, Secretaria de Educação Média e Tecnológica (1999).
Parâmetros Curriculares Nacionais: ensino médio. Brasília: MEC/SEMT.
Brazil, D., Coulthard, M. and Johns, C. (1980) Discourse Intonation and Language Teaching.
London: Longman.
Brazil, D. (1985) The communicative value of intonation in English. English Language
Research, University of Birmingham.
Brazil, D. (1994a) Pronunciation for Advanced Learners of English: Student s Book.
Cambridge: Cambridge University Press.
Brazil, D. (1994b) Pronunciation for Advanced Learners of English: Teacher s Book.
Cambridge: Cambridge University Press.
Bresnahan, M. J. et al. (2002). Attitudinal and affective response toward accented English
Language & Communication. 22: 171-185. www.elsevier.com/locate/lang.com (11
November 2003).
Broughton, G., Brumfit, C., Flavell, R., Hill, P. and Pincas, A. (1980) Teaching English as a
Foreign Language (2nd edition). London: Routledge & Kegan Paul.
Brown, A. (1992) Twenty questions . In A. Brown (ed.) Approaches to Pronunciation
Teaching. London: MacMillan.
Brown G. and Yule G. (1983) Discourse Analysis. Cambridge: Cambridge University Press.
Brown, H. D. (1994) Teaching by Principles. New Jersey: Prentice Hall.
Brown, J. D. (1988) Understanding Research in Second Language Learning: a teacher s
guide to statistics and research design. Cambridge: Cambridge University Press.
72
Callamand, M. (1981) Méthodologie de l enseignement de la prononciation. Paris: Nathan.
Callamand, M. and Pedoya, E. (1984) Phonétique et ensignement . Le français dans le
monde182:56-58.
Cambridge International Dictionary of English (1995) Cambridge: Cambridge University
Press.
Canale, M. (1983) From communicative competence to communicative language pedagogy .
In J. C. Richards and R. W. Schmidt (eds.) Language and Communication. London:
Longman.
Canale, M. and Swain, M. (1980) Theoretical basis of communicative approaches to second
language teaching and testing . Applied Linguistics 1:1-47.
Carruthers, R. (1987) Teaching pronunciation . In M. H. Long and J. C. Richards (eds.)
Methodology in TESOL: a Book of Readings. Boston: Heinle & Heinle Publishers.
Cauldwell, R. and Allan M. (1998) Phonology. Birmingham: The University of Birmingham.
MA TEFL/TESL Open Distance Learning.
Celce-Murcia, M. (1987) Teaching pronunciation as communication . In J. Morley (ed.)
Current Perspectives on Pronunciation. Washington, DC: TESOL
Coulthard, R. M. (1985) An Introduction to Discourse Analysis (2nd edition). London:
Longman.
Couper-Kuhlen, E. (1986) An Introduction to English Prosody. London: Edward Arnold.
Crystal, D. (1995) The Cambridge encyclopaedia of the English language. Cambridge:
Cambridge University Press.
Dickerson, L. (1975) The learner s interlanguage as a set of variable rules . TESOL
Quarterly. 9 (4): 401-408.
Deterding, D. and Poedjosoedarmo, G. (1998) The Sounds of English: Phonetics and
Phonology for English Teachers in Southeast Asia. Singapore: Prentice Hall.
Eckman, F. (1977) Markedness and the Contrastive Analysis Hypothesis . Language
Learning 27: 315-330.
Fillmore, C. J. (1977) Topics in lexical semantics . In R. W. Cole (ed.) Current Issues in
Linguistic Theory. Bloomington: Indiana University Press.
Firth, J. R. (1957) Papers in Linguistics. Oxford: Oxford University Press.
Flege, J. (1987) A Critical Period for Learning to Pronounce Second Languages? Applied
Linguistics 8: 162-177.
73
Francis, A. L. and Jones, E. (1996) Phonetics and Phonological Theory . Language &
Communication 16 (4): 381-393 www.elsevier.com/locate/system (4 January 2004).
Fries, C. (1945) Teaching and Learning English as a Foreign Language. Ann Arbor:
University of Michigan Press.
Gardner, R. C. (1985) Social Psychology and Second Language Learning: The Role of
Attitudes and Motivation. London: Edward Arnold.
Gilbert, J. B. (1993) Clear Speech: Pronunciation and Listening Comprehension in North
American English (2nd edition). New York: Cambridge University Press.
Goto, H. (1971) Auditory Perception by Normal Japanese Adults of the Sounds l and r .
Neuropsychologia 9: 317-323.
Goh, C. C. M. (2000) A cognitive perspective on language learners listening comprehension
problems . System 28: 55-75 www.elsevier.com/locate/system (12 January 2004).
Graham, C. (nd) Contrastive Analysis Revisited: A Participative Approach to Improving
Language Performance . www.unb.br/il/let/graham/conan.htm (14 February 2004).
Harnsberger, D. J. (2001) The perception of Malayalam nasal consonants by Marathi,
Punjabi, Tamil, Oriya, Bengali, an American English listeners: A multidimensional
scaling analsis . Journal of Phonetics 29: 303-327. Available at www.idealibrary.com
(1 February 2004).
Horwitz, E. K. (1999) Cultural and situational influences on foreign language learners
beliefs about language learning: a review of BALLI studies . System 27: 557-576
www.elsevier.com/locate/system (1 February 2004).
Hutchinson, T. (1998) American Hotline. Oxford: Oxford University Press.
Ivani , R. Camps, D. (2001) I am how I sound. Voice as self-representation in L2 writing .
Journal of Second Language Writing 10: 3-33. www.elsevier.com/locate/system
(1 March 2004).
Jackson, C. (2004) www.cindyjackson.com/my_surgery.html (12 March 2004).
Kello, C. T. (2003) Patters of timing in the acquisition, perception, and production of
speech . Journal of Phonetics 31: 619-626. Available at www.sciencedirect.com
(12 December 2004).
Kenworthy, J. (1987) Teaching English Pronunciation. London: Longman.
74
Koren, S. (1995) Foreign Language Pronunciation Testing: a new approach . System 23 (3):
387-400 Elsevier www.elsevier.com/locate/system (10 March 2004).
Krashen, S. D. and Terrell T. D. (1983) The Natural Approach: Language Acquisition in the
classroom. Hayward, California: Alassany Press.
Larsen-Freeman, D. (1986) Techniques and Principles in Language Teaching. New York:
Oxford University Press.
Leather, J. and James, A. (1991) The Acquisition of Second Language Speech . Studies in
Second Language Acquisition 13: 305-331.
Long, M. H. (1983) Does second language instruction make a difference? A review of
research . TESOL Quarterly 17(3): 359-382.
Maken, M. A. and Ferguson, C. A. (1987) Phonological Universals in Language
Acquisition . In G. Ioup and S. H. Weinberger (eds.) Interlanguage Phonology.
Cambridge: Newbury House.
Martins, M. H. (1982) O que é Leitura. São Paulo: Brasiliense.
McCarthy, M. (1991) Discourse Analysis for Language Teachers. Cambridge: Cambridge
University Press.
Miller, L. M. (2004) Using learning styles to evaluate computer-based instruction .
Computers in Human Behavior. In press. Available online at www.sciencedirect.com
(19 March 2004)
Naiman, N. (1992) A communicative approach to pronunciation teaching . In P. Avery and
S. Ehrlich, Teaching American English Pronunciation. Oxford: Oxford University
Press.
Newmark, L. (1990) Changing Student Pronunciation: Getting People to Eat Worms . In R.
C. Scarcella, E. S. Andersen and S. D. Krashen (eds.) Developing Communicative
Competence in a Second Language. Boston, MA.: Heinle & Heinle Publishers.
Newton, Anne C. (1979) Current Trends in Language Teaching . In M. Celce-Murcia & L.
McIntosh (eds.) Teaching English as a second or foreign language. Rowley, MA:
Newbury House.
Nunan, D. (1991) Language Teaching Methodology: A textbook for teachers. Hemel
Hempstead: Prentice Hall International.
O Connor, J. D. and Fletcher, C. (1989) Sounds English. Harlow: Longman.
Ohala, J. J. and Shriberg, E. E. (1990) Hypercorrection in speech perception . Proceedings of
the XIIth International Congress of Phonetic Sciences. Aix-en-Provence 1 : 1-16.
75
Oxenden, C. and Seligson, P. (1996) English File: student s book one. Oxford: Oxford University
Press.
Pavlov, I. P. (1955) Selected works. (J. Gibbons, ed.). Moscow: Foreign Languages Pub.
House.
Pennington, M. C., and Richards, J. C. (1986) Pronunciation revisited . TESOL Quarterly
20(2): 207-225.
Pennington, M. C. (1996) Phonology in English Language Teaching: An International
Approach. London: Longman.
Pica, T. (2000) Tradition and transition in English language teaching methodology . System
28: 1-18 www.elsevier.com/locate/system (12 January 2004).
Pica, T. (1984) Pronunciation Activities with an Accent on Communication . English
Teaching Forum 22(3): 2-6.
Pienemann, M. (1984) Psychological Constraints on the Teachability of Languages . Studies
in Second Language Acquisition 6:186-214.
Poedjosoedarmo, G. (2004) O Ensino da Pronúncia: por quê, o quê, quando e como. São
Paulo: Portfolio SBS: reflexões sobre o ensino de idiomas; 9. SBS Editora.
Rogerson, P. and Gilbert J. B. (1990) Speaking Clearly. Cambridge: Cambridge University
Press.
Rossiter, M. J. and Derwing, T. M. (2002) ESL learners perceptions of their pronunciation
needs and strategies . System 30: 155-166. Available at www.elsevier.com/locate/system
(20 February, 2002).
Rutherford, W. E. (1987) Second Language Grammar: Learning and Teaching. New York:
Longman.
Sadock, J. M. (1978) On testing for conversational implicature . In P. Cole (ed.) Syntax and
Semantics, Volume 9: Pragmatics. New York: Academic Press.
Sato, C. (1985) Task variation in interlanguage phonology . In S. Gass and E. Varonis (eds.)
Input in Second Language Acquisition. Rowley, MA.: Newbury House.
Scarcella, R. C. and Oxford, R. L. (1994) Second Language Pronunciation: State of the Art in
Instruction. System 22(2): 221-230.
Schank, R. C. (1979) Some prerequisites for a computational pragmatics . In J. L. Moy (ed.)
Pragmalinguistics. The Hague: Mouton.
Seliger, H. W. and Shohamy, E. (1989) Second Language Research Methods. Oxford: Oxford
University Press.
76
Shedivy, S. L. (2004) Factors that lead some students to continue the study of foreign
language past the usual 2 years in high school . System 32:103-119. Elsevier
www.elsevier.com/locate/system (10 March 2004).
Sinclair, J. McH. and Brazil, D. (1982) Teacher Talk. Oxford: Oxford University Press.
Skinner, B. F. (1957) Verbal Behaviour. New York: Appleton Century Crofts.
Stern, H. H. (1992) Issues and Options in Language Teaching. Oxford: Oxford University
Press.
Tarone, E. (1978) The Phonology of Interlanguage . In J. C. Richards (ed.) Understanding
Second and Foreign Language Learning. Rowley, MA: Newbury House.
The Concise Oxford Dictionary (8th edition) (1990). Oxford: Oxford University Press.
Underhill, A. (1994) Sound Foundations. Oxford: Macmillan Heinemann.
Ur, P. (1981) Discussions that Work. Cambridge: Cambridge University Press.
UCLES
University of Cambridge Local Examinations Syndicate (1991) Cambridge
Proficiency Examination Practice 4: Teacher s Book. Cambridge: Cambridge
University Press.
Woods, A., Fletcher, P, and Hughes, A. (1986) Statistics in language studies. Cambridge:
Cambridge University Press.
Wrembel, M. (2001) Innovative Approaches to the Teaching of Practical Phonetics . Adam
Michiewicz University. www.phon.ucl.ac.uk/home/johnm/ptlc2001/pdf/wrembel.pdf.
(5 February 2004).
77
This document was created with Win2PDF available at http://www.daneprairie.com.
The unregistered version of Win2PDF is for evaluation or non-commercial use only.